{"id": "2603.16648", "title": "Domain-Independent Dynamic Programming with Constraint Propagation", "abstract": "There are two prevalent model-based paradigms for combinatorial problems: 1) state-based representations, such as heuristic search, dynamic programming (DP), and decision diagrams, and 2) constraint and domain-based representations, such as constraint programming (CP), (mixed-)integer programming, and Boolean satisfiability. In this paper, we bridge the gap between the DP and CP paradigms by integrating constraint propagation into DP, enabling a DP solver to prune states and transitions using constraint propagation. To this end, we implement constraint propagation using a general-purpose CP solver in the Domain-Independent Dynamic Programming framework and evaluate using heuristic search on three combinatorial optimisation problems: Single Machine Scheduling with Time Windows, the Resource Constrained Project Scheduling Problem (RCPSP), and the Travelling Salesperson Problem with Time Windows (TSPTW). Our evaluation shows that constraint propagation significantly reduces the number of state expansions, causing our approach to solve more instances than a DP solver for Single Machine Scheduling and RCPSP, and showing similar improvements for tightly constrained TSPTW instances. The runtime performance indicates that the benefits of propagation outweigh the overhead for constrained instances, but that further work into reducing propagation overhead could improve performance further. Our work is a key step in understanding the value of constraint propagation in DP solvers, providing a model-based approach to integrating DP and CP.", "date": "2026-03-17", "categories": []} {"id": "2606.19533", "title": "A Tool for the Synthesis of Adaptive Probabilistic Processors Based on the Ising Model", "abstract": "This work presents a tool for the synthesis and simulation of probabilistic architectures for solving combinatorial optimization problems by mapping them to the Ising model. The proposed approach automatically constructs the Ising Hamiltonian and determines the number of probabilistic elements (p-bits) based on problem characteristics such as size and topology. Furthermore, the tool introduces an adaptive strategy for selecting the most suitable update algorithm among Gibbs Sampling, Simulated Annealing (SA), Simulated Quantum Annealing (SQA), and cluster-based methods. Experimental results using benchmark problems demonstrate improved convergence behavior and flexibility compared to fixed approaches. The proposed framework enables systematic evaluation of probabilistic computing strategies and supports the development of future hardware implementations based on MTJs and p-bits.", "date": "2026-06-17", "categories": []} {"id": "2606.19741", "title": "Interpreting Neural Combinatorial Optimization via Evolving Programmatic Bottlenecks", "abstract": "Neural Combinatorial Optimization (NCO) achieves strong performance, yet its black-box nature remains a key roadblock to deployment and scientific diagnosis. Standard interpretability tools, such as Concept Bottleneck Models (CBMs), are ill-equipped for NCO, whose decisions are dynamic, state-dependent, and lack proper concept vocabulary definition. To close this gap, we introduce Evolving Programmatic Bottlenecks (EPB), to our knowledge, the first framework for interpreting NCO policies by distilling black-box NCO models into human-readable program portfolios. EPB employs an LLM to autonomously evolve a bank of programs, where each program's per-step action distribution serves as the bottleneck. EPB works through an iterative framework: Block I fixes program bank capacity and introduces a hybrid textual-numerical gradient descent scheme that couples numerical gradients for student router updates and textual gradients for LLM-based program revision; Block II dynamically adapts bank capacity via fault-targeted expansion and redundancy pruning. Extensive experiments demonstrate EPB's effectiveness and broad applicability, where the distilled program portfolios largely match original performance. EPB also reveals that NCO behavior shifts across optimization stages and can be approximated as a composition of classic heuristic variants. Our work advances interpretable NCO and establishes EPB as a promising tool for interpreting sequential decision-making models.", "date": "2026-06-18", "categories": []} {"id": "2606.20142", "title": "RACL: Reasoning-Agent Control Layers for Continuous Metaheuristic Learning", "abstract": "This paper introduces RACL, a Reasoning-Agent Control Layer for metaheuristics. RACL places a reasoning agent above an existing optimizer. The agent does not replace the optimizer and does not modify business constraints. Instead, it controls the optimizer's internal search behavior by observing operational memory, reasoning over past behavior, formulating bounded hypotheses, testing interventions, evaluating outcomes, applying guardrails, consolidating useful policies and explaining its decisions. The experiment uses vehicle routing as a testbed, but the contribution is not a new routing solver, a particular ALNS configuration or a specific set of routing rules. The contribution is the RACL method: a way for a reasoning agent to discover, validate, consolidate and explain algorithmic control rules for a metaheuristic. In the current experimental setting, RACL improves or ties the Operational Memory Policy in 21 of 21 feasible cases and improves or ties a non-reasoning Stagnation-Triggered Policy in 18 of 21 feasible cases, with an average RACL vs STP cost delta of -0.641%. In the Sevilla-9/10 runtime sample, RACL improves average cost by -8.337% versus Fixed and -1.605% versus STP without showing material computational overhead. During the proof-of-concept, Codex was used as an in-the-loop reasoning agent observing executions, interpreting logs and proposing live bounded interventions. The policy proxy was later used only to make quantitative evaluation reproducible.", "date": "2026-06-18", "categories": []} {"id": "2508.20330", "title": "FORGE: Foundational Optimization Representations from Graph Embeddings", "abstract": "Combinatorial optimization problems are ubiquitous in science and engineering. Still, learning-based approaches to accelerate combinatorial optimization often require solving a large number of difficult instances to collect training data, incurring significant computational cost. Existing learning-based methods require training dedicated models for each problem distribution, for each downstream task, severely limiting their scalability and generalization. We introduce Forge: Foundational Optimization Representations from Graph Embeddings, a framework that pre-trains a vector-quantized graph autoencoder on a large, diverse collection of mixed-integer programming (MIP) instances in an unsupervised manner, without relying on optimization solvers or optimal solutions. Vector quantization produces discrete code assignments that serve as a vocabulary for representing optimization instances. We evaluate Forge in both unsupervised and supervised settings. In the unsupervised setting, Forge embeddings effectively cluster unseen instances across problem domains and sizes. In the supervised setting, we fine-tune Forge embeddings and show that a single pre-trained model helps predicting both the integrality gap for cut-generation and variable hints for search guidance across multiple problem and size distributions. In both tasks, we improve the performance of a commercial optimization solver and outperform state-of-the-art learning-based methods. Finally, we open-source our training code, pre-trained Forge weights, and embeddings for multiple MIP distributions to foster further research in representation learning for optimization problems https://skadio.github.io/forge/", "date": "2025-08-28", "categories": []} {"id": "2510.27353", "title": "An In-depth Study of LLM Contributions to the Bin Packing Problem", "abstract": "Recent studies have suggested that Large Language Models (LLMs) could provide interesting ideas contributing to mathematical discovery. This claim was motivated by reports that LLM-based genetic algorithms produced heuristics offering new insights into the online bin packing problem under uniform and Weibull distributions. In this work, we reassess this claim through a detailed analysis of the heuristics produced by LLMs, examining both their behavior and interpretability. Despite being human-readable, these heuristics remain largely opaque even to domain experts. Building on this analysis, we propose a new class of algorithms tailored to these specific bin packing instances. The derived algorithms are significantly simpler, more efficient, more interpretable, and more generalizable, suggesting that the considered instances are themselves relatively simple. We then discuss the limitations of the claim regarding LLMs' contribution to this problem, which appears to rest on the mistaken assumption that the instances had previously been studied. Our findings instead emphasize the need for rigorous validation and contextualization when assessing the scientific value of LLM-generated outputs.", "date": "2025-10-31", "categories": []} {"id": "2602.23092", "title": "Enhancing CVRP Solver through LLM-driven Automatic Heuristic Design", "abstract": "The Capacitated Vehicle Routing Problem (CVRP), a fundamental combinatorial optimization challenge, focuses on optimizing fleet operations under vehicle capacity constraints. While extensively studied in operational research, the NP-hard nature of CVRP continues to pose significant computational challenges, particularly for large-scale instances. This study presents AILS-AHD (Adaptive Iterated Local Search with Automatic Heuristic Design), a novel approach that leverages Large Language Models (LLMs) to revolutionize CVRP solving. Our methodology integrates an evolutionary search framework with LLMs to dynamically generate and optimize ruin heuristics within the AILS method. Additionally, we introduce an LLM-based acceleration mechanism to enhance computational efficiency. Comprehensive experimental evaluations against state-of-the-art solvers, including AILS-II and HGS, demonstrate the superior performance of AILS-AHD across both moderate and large-scale instances. Notably, our approach establishes new best-known solutions for 8 out of 10 instances in the CVRPLib large-scale benchmark, underscoring the potential of LLM-driven heuristic design in advancing the field of vehicle routing optimization.", "date": "2026-02-26", "categories": []} {"id": "2605.30920", "title": "Unsupervised Diffusion Solver for Combinatorial Optimization via Combinatorial Adjoint Matching", "abstract": "Diffusion-based neural solvers have shown strong promise for combinatorial optimization (CO), but existing methods typically rely on supervised training with large collections of near-optimal solutions. In this work, we extend adjoint-based trajectory optimization methods to discrete combinatorial domains. We formulate diffusion-based CO as a stochastic control problem over Continuous-Time Markov Chains and introduce discrete adjoint dynamics for propagating optimization signals through discrete generative trajectories. Building on this formulation, we propose Combinatorial Adjoint Matching (CAM), an unsupervised training framework for discrete diffusion solvers with structured and low-variance trajectory-level optimization signals. Empirically, CAM consistently outperforms existing unsupervised diffusion baselines and achieves performance competitive with strong supervised diffusion solvers and even traditional solvers across diverse combinatorial optimization problems. Our code is available at https://github.com/Shengyu-Feng/CAM.", "date": "2026-05-29", "categories": []} {"id": "2606.18105", "title": "OmniPlan: An Adaptive Framework for Timely and Near-Optimal Network Planning Optimization", "abstract": "Network planning optimization is a fundamental problem across diverse domains, including transportation systems, communication networks, and power grids. It requires simultaneous optimization of multiple competing objectives under complex constraints. Existing network planning optimization frameworks rely on mixed integer programming (MIP) solvers, heuristics, and deep reinforcement learning (DRL) models to compute planning decisions. However, they lack effective adaptability to diverse and dynamic user intents, thus leading to the trade-off between execution time and optimality. In this paper, we propose OmniPlan, an adaptive framework that achieves both timeliness and near-optimality in network planning optimization. To achieve the adaptability lacking in existing solutions, OmniPlan employs a large language model (LLM)-based interpreter to convert heterogeneous natural-language intents into a unified and quantifiable user-preference vector. Then it employs a mixture-of-experts architecture that integrates MIP solvers, heuristics, and DRL models as specialized experts, where OmniPlan adapts to diverse intents by dynamically selecting timely and near-optimal experts. Finally, it incorporates a DRL-based expert configuration module that fine-tunes optimization objective weights to align planning decisions with user-specific preferences. We evaluate OmniPlan with a representative real-world workload, i.e., distributed machine learning (ML), where we leverage OmniPlan to offload a wide spectrum of ML inference tasks, e.g., decision trees, SVM, naive Bayes, XGBoost, and random forests, onto a network of hardware devices. Our experiments on a real-world testbed indicate that OmniPlan achieves near-optimal and low-execution-time offloading for real-world ML inference tasks, reducing latency by up to 97.8\\% and network device resource consumption by up to 11.5\\%.", "date": "2026-06-16", "categories": []} {"id": "2606.18514", "title": "N(CO)$^2$: Neural Combinatorial Optimization with Chance Constraints to Solve Stochastic Orienteering", "abstract": "Neural combinatorial optimization (NCO) offers a promising alternative to traditional heuristic-based methods for solving complex graph optimization problems by proposing to learn heuristics through data. This class of problems frequently arises in automation, as it can be used to model a variety of applications. While NCO has been extensively studied for deterministic combinatorial optimization problems, there are only a few works that aim to solve stochastic combinatorial optimization problems. In this work, we present N(CO)$^2$: Neural Combinatorial Optimization with Chance cOnstraints to solve the Stochastic Orienteering Problem (SOP) without the use of hand-crafted heuristics. By integrating a reinforcement learning (RL) framework, the model optimizes path selection under uncertainty, effectively balancing exploration and exploitation. Empirical results demonstrate that our method generalizes well across diverse SOP instances, achieving competitive performance compared to the state-of-the-art mixed-integer linear program (MILP) for the task. The proposed approach reduces human effort in heuristic design while enabling adaptive and efficient decision-making in uncertain environments.", "date": "2026-06-16", "categories": []} {"id": "2606.18730", "title": "Two-Phase Bilevel Search for the Moving-Target Traveling Salesman Problem with Moving Obstacles", "abstract": "The Moving-Target Traveling Salesman Problem (MT-TSP) seeks a minimum cost trajectory for an agent that departs from a static depot, visits a set of moving targets, each within one of their assigned time windows, and returns to the depot. In this article, we study the Moving-Target Traveling Salesman Problem with Moving Obstacles (MT-TSP-MO), a generalization of the MT-TSP where the agent trajectory must avoid moving obstacles. We present a Mixed-Integer Conic Programming (MICP) formulation that can be solved using off-the-shelf solvers, as well as a fast and scalable Two-Phase Bilevel Search (TPBS) algorithm that computes high-quality feasible solutions for the problem. We evaluate our approaches against an existing baseline algorithm on a broad range of problem instances with up to 40 targets and 40 obstacles. The results demonstrate that both the proposed methods significantly outperform the baseline with respect to success rates, solution costs, and computation time.", "date": "2026-06-17", "categories": []} {"id": "2606.18807", "title": "Learning Augmented Exact Exponential Algorithms", "abstract": "The field of learning-augmented algorithms has demonstrated that machine-learned predictions can bypass worst-case lower bounds across a wide range of problems. So far, however, the focus has been almost exclusively on polynomial-time algorithms, where predictions improve competitive ratios, approximation guarantees, or running times. In this paper, we raise the question of whether predictions can push the frontier of exact exponential-time algorithms for NP-hard problems. We answer this question affirmatively by proposing a general approach that augments an entire family of state-of-the-art exact algorithms for a variety of subset selection problems. We show that a noisy predictor that is only marginally better than random guessing suffices to provably reduce the search space, and that the resulting runtime speedup scales smoothly with the prediction quality. Importantly, our algorithms require only pairwise independence of predictions or, alternatively, do not require the knowledge of the predictor's accuracy - both strictly weaker and more realistic settings than typically assumed.", "date": "2026-06-17", "categories": []} {"id": "2606.19185", "title": "AGDN: Learning to Solve Traveling Salesman Problem with Anisotropic Graph Diffusion Network", "abstract": "The Traveling Salesman Problem (TSP) is a cornerstone of combinatorial optimization and arises in many practical scenarios. Although graph-based learning approaches have been explored for TSP, the question of how to exploit graph structure more effectively remains open. We present the Anisotropic Graph Diffusion Network (AGDN), a new Graph Neural Network framework designed to solve TSP. Our method tackles two central difficulties: (1) the lack of informative topological prior in fully connected TSP graphs, and (2) losing connected nodes in the optimal solution after the commonly used graph sparsification techniques. To overcome these issues, we construct a MixScore transition matrix that merges node similarity with pairwise distance, and we develop an anisotropic graph diffusion strategy that supports efficient information exchange across multiple hops. Comprehensive experiments spanning diverse instance sizes and node distributions show that AGDN consistently outperforms existing methods while keeping computation time competitive. Furthermore, AGDN generalizes well to problem sizes and distributions beyond those seen during training. The implementation is publicly available at: https://github.com/LabRAI/AGDN.", "date": "2026-06-17", "categories": []} {"id": "2505.13986", "title": "RIDGECUT: Learning Graph Partitioning with Rings and Wedges", "abstract": "Reinforcement learning (RL) has shown promise for combinatorial optimization problems on graphs by learning heuristics that generalize across instances. However, effectively incorporating domain knowledge into RL frameworks for graph partitioning remains challenging, as existing approaches typically rely on unconstrained node-level actions that lead to large action spaces and inefficient exploration. In this paper, we propose RidgeCut, an RL framework that constrains the action space to enforce structure-aware partitioning in the Normalized Cut problem. Using transportation networks as a motivating example, we introduce a novel concept that leverages domain knowledge about urban road topology -- where natural partitions often take the form of concentric rings and radial wedges. By transforming the graph into linear or circular representations, our method enables the use of transformer-based policies and efficient learning via Proximal Policy Optimization. The resulting partitions from RidgeCut are not only aligned with expected spatial layouts but also achieve lower normalized cuts compared to existing methods. Experimental results on synthetic and real-world traffic graphs demonstrate that RidgeCut consistently outperforms existing methods while exhibiting strong inductive generalization across graph sizes. Although motivated by road networks, RidgeCut provides a general mechanism for embedding structural priors into RL frameworks for graph partitioning.", "date": "2025-05-20", "categories": []} {"id": "2602.08210", "title": "CADO: From Imitation to Cost Minimization for Heatmap-based Solvers in Combinatorial Optimization", "abstract": "Heatmap-based solvers have emerged as a promising paradigm for Combinatorial Optimization (CO). However, we argue that the dominant Supervised Learning (SL) training paradigm suffers from a fundamental objective mismatch: minimizing imitation loss (e.g., cross-entropy) does not guarantee solution cost minimization. We dissect this mismatch into two deficiencies: Decoder-Blindness (being oblivious to the non-differentiable decoding process) and Cost-Blindness (prioritizing structural imitation over solution quality). We empirically demonstrate that these intrinsic flaws impose a hard performance ceiling. To overcome this limitation, we propose CADO (Cost-Aware Diffusion models for Optimization), a streamlined Reinforcement Learning fine-tuning framework that formulates the diffusion denoising process as an MDP to directly optimize the post-decoded solution cost. We introduce Label-Centered Reward, which repurposes ground-truth labels as unbiased baselines rather than imitation targets, and Hybrid Fine-Tuning for parameter-efficient adaptation. CADO achieves state-of-the-art performance across diverse benchmarks, validating that objective alignment is essential for unlocking the full potential of heatmap-based solvers.", "date": "2026-02-09", "categories": []} {"id": "2606.14764", "title": "Avoiding Exponential Blow-Up in Distributive Lattice Submodular Minimization", "abstract": "Submodular function minimization has gained a lot of interest in recent years. They are highly applicable in the area of Computer Vision and Machine Learning. Often such applications require to work with submodular functions defined on distributive lattice. Current best way of dealing with it is using a transformation which extrapolates the submodular function for the respective boolean lattice. It makes optimization system too inefficient due to enlargement of the working space. Quantitatively, the expanded space has additional exponential (in set size) number of elements. We propose a generic framework for dealing with distributive lattice which only works within distributive lattice. Our framework allows one to use already established submodular function minimization algorithms for boolean lattice. In our experiment, we show the huge improvement in terms of running time over tranditional methods for handling distributive lattice.", "date": "2026-06-08", "categories": []} {"id": "2507.10834", "title": "From Small to Large: A Graph Convolutional Network Approach for Solving Assortment Optimization Problems", "abstract": "Assortment optimization seeks to select a subset of substitutable products, subject to constraints, to maximize expected revenue. The problem is NP-hard due to its combinatorial and nonlinear nature and arises frequently in industries such as e-commerce, where platforms must solve thousands of such problems each minute. We propose a graph convolutional network (GCN) framework to efficiently solve constrained assortment optimization problems. Our approach constructs a graph representation of the problem, trains a GCN to learn the mapping from problem parameters to optimal assortments, and develops three inference policies based on the GCN's output. Owing to the GCN's ability to generalize across instance sizes, patterns learned from small-scale samples can be transferred to large-scale problems. Theoretical results are established to show the expressive power of the proposed GCN, and explain the underlying mechanism of the size generalization ability. Numerical experiments show that a GCN trained on instances with 20 products achieves over 85% of the optimal revenue on problems with up to 2,000 products within seconds, outperforming existing heuristics in both accuracy and efficiency. We further extend the framework to settings with an unknown choice model using transaction data and demonstrate similar performance and scalability.", "date": "2025-07-14", "categories": []} {"id": "2604.23841", "title": "Scalable Production Scheduling: Linear Complexity via Unified Homogeneous Graphs", "abstract": "Efficiently solving the Job Shop Scheduling Problem in real-world industrial applications requires policies that are both computationally lean and topologically robust. While Reinforcement Learning has shown potential in automating dispatching rules, existing models often struggle with a scalability bottleneck caused by quadratic graph complexity or the architectural overhead of heterogeneous layers. We introduce a unified graph framework that employs feature-based homogenization to project distinct node roles into a shared latent space. This allows a standard homogeneous Graph Isomorphism Network to capture complex resource contention with linear complexity, ensuring low-latency inference for large-scale industrial applications. Our empirical results demonstrate that our framework achieves state-of-the-art performance while exhibiting consistent zero-shot generalization. We identify the job-to-machine ratio as the primary driver of policy effectiveness, rather than absolute problem size. Based on this, we propose a hypothesis of structural saturation, demonstrating that policies trained on critically congested instances ($\\mathcal{J} \\approx \\mathcal{M}$) learn scale-invariant resolution strategies. Agents trained at this saturation point internalize invariant conflict-resolution logic, allowing them to treat massive rectangular instances as a sequential concatenation of saturated sub-problems. This approach eliminates the need for expensive scale-specific retraining and prevents overfitting to statistical shortcuts, providing a robust and efficient pathway for deploying RL solutions in dynamic production environments.", "date": "2026-04-26", "categories": []} {"id": "2606.13682", "title": "A Deep Reinforcement Learning (DRL)-Based Transformer Method for Solving the Open Shop Scheduling Problem", "abstract": "The open shop scheduling problem (OSSP) arises in many industrial and service settings but remains computationally challenging as the number of jobs and machines increases. While exact methods quickly become intractable, classical dispatching rules and metaheuristics may require substantial tuning to maintain solution quality at large scales. This study develops a Transformer-based scheduling policy for OSSP using an encoder-decoder architecture with multi-head attention. The model is trained on Taillard benchmark instances (4x4, 5x5, 7x7, and 10x10) using only the processing-time matrix as input and produces feasible schedules with makespans typically within 15-30% of best-known values. To evaluate scalability, the trained policy is applied without retraining to randomly generated instances from 40x40 to 100x100 and compared against classical dispatching heuristics, including SPT, LPT, MWKR, and EST. Across these large instances, the Transformer achieved average gaps of 12.89-15.12% relative to a standard lower bound. Compared with EST, the Transformer remained competitive, typically within a modest margin, while substantially outperforming SPT and LPT. These results indicate that a Transformer policy trained on small OSSP instances can generalize to substantially larger problems and provide a feature-light, learning-based alternative to classical dispatching rules.", "date": "2026-03-25", "categories": []} {"id": "2606.14582", "title": "A Temporal Planning Framework for Disruption Aware Dynamic Route Optimization in Heterogeneous Railway Systems", "abstract": "Efficient route optimization play a vital role in ensuring both safety and punctuality in railway operations. It is very crucial particularly in heterogeneous multi-gauge railway networks with varying train speed, stopping pattern, infrastructure compatibility constraints increase coordination complexity. In single-track systems these challenges are further intensify due to all trains to share the same track and requires frequent track switching.Stochastic disruptions events including blocked tracks, blocked trains, engine failure and speed slowdowns introduces additional unpredictability in operations and deviate the timetable. However, existing studies predominantly focuses on high-level timetabling, omitting operational details such as track switching coordination. As a result leaving decision to human operators, increasing safety risks into railway operations. This study proposes a framework based on temporal planning for dynamic route optimization and disruption management in heterogeneous railway systems. The framework formulates railway operations as a temporal planning problem using PDDL 2.1 with explicitly modeling gauge compatibility constraints and diverse disruption scenarios. It generates conflict-free timestamped operational plans specifying both optimized schedules and executable action sequences. To evaluate the proposed framework, we developed a benchmark problem set with 200 instances using up to 1,000 track points and 120 trains. Two state-of-the-art temporal planners and a plan validator were employed to assessed the framework. The experimental results demonstrate that the framework effectively generates temporal operational plans for heterogeneous railway systems and handles multi-gauge constraints, disruptions, and reduces dependence on manual decision making.", "date": "2026-06-12", "categories": []} {"id": "2604.20236", "title": "Machine Learning-based Two-Stage Graph Sparsification for the Travelling Salesman Problem", "abstract": "High-performance TSP solvers such as Lin-Kernighan-Helsgaun (LKH) search within a \\emph{candidate graph} -- a small subset of edges pre-selected for the solver -- rather than over the complete graph. The two leading sparsification heuristics, $\\alpha$-Nearest and POPMUSIC, each fall short of the density-coverage balance: $\\alpha$-Nearest is dense with stable recall, while POPMUSIC is sparser but its recall degrades with scale. Their union closes the recall gap while remaining far below the complete graph in density, leaving room for further reduction. Existing learning-based sparsifiers score edges on the complete graph, an approach that is expensive and largely limited to Euclidean instances. We propose a two-stage method that inverts this logic. Stage~1 takes the union of $\\alpha$-Nearest and POPMUSIC, achieving near-perfect recall at ${\\sim}6N$ edges. Crucially, the union annotates each edge with its \\emph{source provenance} -- whether it was endorsed by $\\alpha$-Nearest, POPMUSIC, or both. Stage~2 trains a lightweight classifier on these annotated edges and prunes the lowest-scoring ones. Because dual-source edges are almost always optimal, the learning problem reduces to filtering the single-source subset -- a substantially easier task than classifying all $O(N^2)$ edges from scratch. Across four distance types, five spatial distributions, and problem sizes from 50 to 500, the pipeline reduces candidate-graph density by $37$-$47\\%$ while retaining ${\\geq}99.69\\%$ of optimal-tour edges, and matches or exceeds the coverage of recent Euclidean-only neural sparsifiers at lower density at TSP500.", "date": "2026-04-22", "categories": []} {"id": "2606.13407", "title": "Optimizing Appliance Scheduling for Solar Energy Management Using Metaheuristic Algorithms", "abstract": "Renewable energy is essential for meeting future energy demands; however, solar energy generation, which occurs only during daylight hours often does not align with household consumption patterns. Appliances such as cookers, washing machines, and dryers are typically operated according to user preferred schedules rather than solar energy availability, creating a scheduling optimization problem. The objective is to determine optimal appliance start times to maximize renewable energy utilization while minimizing user inconvenience and adhering to system constraints. This paper presents a metaheuristic approach using Iterated Local Search (ILS) and Simulated Annealing (SA) to optimize appliance start times, while considering appliance operating durations, power consumption, inverter limit, battery state of charge constraints, and solar generation forecasts. Unlike most existing work, the scheduling is extended beyond a single day to accommodate unfinished tasks from previous days (spillover), ensuring operational continuity and enabling sequential operation across multiple days. Experimental results show that the sequential multi-day scheduling framework effectively manages system constraints while ensuring user convenience under exclusive solar generation. These findings also open opportunities for future research on multi-objective trade-offs between investment in equipment of various sizes, return on that investment, and user satisfaction.", "date": "2026-06-11", "categories": []} {"id": "2601.04884", "title": "Precomputing Multi-Agent Path Replanning Using Temporal Flexibility", "abstract": "Executing a multi-agent plan can be challenging when an agent is delayed, because this typically creates conflicts with other agents. So, we need to quickly find a new safe plan. Replanning only the delayed agent often does not yield an efficient plan, and sometimes cannot even yield a feasible one. On the other hand, replanning other agents may lead to a cascade of changes and delays, and it is computationally expensive. We show how to efficiently replan a single delayed agent by tracking and using the temporal flexibility of other agents while avoiding cascading delays. This flexibility is the maximum delay that the agent can take without changing the order with agents other than the initially delayed agent, or further delaying other agents. Our algorithm, FlexSIPP, precomputes all possible plans for the delayed agent and returns the changes to the other agents within the given scenario. We demonstrate our method in a real-world case study of replanning trains in the densely-used Dutch railway network and in the MovingAI MAPF benchmark set. Our experiments show that FlexSIPP provides effective solutions relevant to real-world adjustments, and within a reasonable timeframe.", "date": "2026-01-08", "categories": []} {"id": "2606.12279", "title": "Mathematical perspective on genetic algorithms with optimization guided operators", "abstract": "Recent work in ML applies genetic algorithms at inference time to iteratively improve solutions to optimization problems. The basic mutation and recombination operators involved are qualitatively different from those studied classically. Mutations are no longer random; an ML algorithm mutates a solution with the goal of improving an objective. Similarly, recombination is not based on random collages of parent solutions. Instead, it is an ML optimization-based operator whose goal is to synthesize improved solutions from its inputs. Thus, these mutation and recombination operators are more likely to improve the objective, but their computational cost is much higher.\n We introduce a general model of genetic algorithms and formulating optimization in this model as a query-complexity problem, using the language of reinforcement learning. We then study specialized models. We show that some optimization problems require generation, mutation, and recombination to be solved. We then obtain qualitatively tight algorithms for a family of problems within this framework that captures the nontrivial role of diversity in the solution pool, a key feature of practical ML genetic algorithms.", "date": "2026-06-10", "categories": []} {"id": "2506.03672", "title": "Latent Guided Sampling for Combinatorial Optimization", "abstract": "Combinatorial Optimization problems are widespread in domains such as logistics, manufacturing, and drug discovery, yet their NP-hard nature makes them computationally challenging. Recent Neural Combinatorial Optimization (NCO) methods leverage deep learning to learn policies for constructing solutions, trained via Supervised or Reinforcement Learning. While promising, these approaches often rely on task-specific augmentations, perform poorly on out-of-distribution instances, and lack robust inference mechanisms. Moreover, existing latent space models either require labeled data or use an instance-independent latent distribution. In this work, we propose LGS-Net, a novel latent space model that conditions on problem instances, and introduce an efficient inference method, Latent Guided Sampling (LGS), based on Markov Chain Monte Carlo and Stochastic Approximation. We show that the iterations of our method form a time-inhomogeneous Markov Chain and provide rigorous theoretical convergence guarantees. Empirical results on benchmark routing tasks show that our method achieves state-of-the-art performance among NCO baselines.", "date": "2025-06-04", "categories": []} {"id": "2601.20970", "title": "The hyper-scaled NLP bound for maximum-entropy remote sampling", "abstract": "The maximum-entropy remote sampling problem (MERSP) is to select a subset of $s$ random variables from a set of $n$ random variables, so as to maximize the information concerning a set of target random variables that are not directly observable. We assume that the set of all of these random variables follows a joint Gaussian distribution, and that we have the covariance matrix available. Finally, we measure information using Shannon's differential entropy.\n The main approach for exact solution of moderate-sized instances of MERSP has been branch-and-bound (B\\&B), and so previous work concentrated on upper bounds. Prior to our work, there were two upper-bounding methods for MERSP: the so-called ``complementary NLP bound'' and the ``spectral bound'', both introduced 25 years ago. We are able now to establish domination results between these two upper bounds. Further, we propose a novel and effective ``hyper-scaled NLP bound'' (hNLP bound) based on a subtle convex relaxation. The ``complementary'' version of hNLP bound for MERSP generalizes the previous complementary NLP bound for MERSP. We provide theoretical guarantees, giving sufficient conditions under which the complementary hNLP bound strictly dominates the complementary NLP bound. In addition, the hNLP formulation allows us to derive upper bounds for rank-deficient covariance matrices when they satisfy a technical condition. This is in contrast to the previous NLP bound that worked with only positive definite covariance matrices (because it was wedded to a complementary formulation). Additionally, we describe procedures for calculating hyper-scaling parameters. Finally, for B\\&B, we provide a variable-fixing methodology and results guiding the best way to construct subproblems. Numerical experiments on benchmark instances demonstrate the effectiveness of our approaches in advancing the algorithmic state-of-the-art for MERSP.", "date": "2026-01-28", "categories": []} {"id": "2606.03419", "title": "Optimizing Explicit Unit-Distance Lower-Bound Certificates", "abstract": "The 2026 disproof of Erd\\H{o}s's unit-distance conjecture and Sawin's quantitative refinement show that the maximum number $u(n)$ of unit distances among $n$ planar points can exceed $n^{1+\\varepsilon}$ for a fixed positive $\\varepsilon$. Sawin's explicit bound gives more than $n^{1.014}$ unit distances for arbitrarily large $n$ and exposes integer parameters whose choice is not fully optimized. This report treats Sawin's parameter selection as a nonlinear integer optimization problem and develops an open-source Python optimization and verification pipeline for certificates involving prime sets $T$ and $S_Q$, integer multiplicities $k(p)$, and a rationally encoded real parameter $R$. After reproducing Sawin's certificate with $\\delta=0.014114\\ldots$, the pipeline yields improved certificates with the same $T$. We develop a tailored integer evolution strategy achieving a certificate with $\\delta=0.015263\\ldots$ and supporting the cautious statement $u(n)>n^{1.0152}$ for arbitrarily large $n$. For extended ramified prime ranges, the Emmerich--Cordella certificate obtained with the same framework reports $u(n)>n^{1.031}$ for $\\#T=67$, illustrating the importance of enlarging $T$. Very recent MathOverflow discussions, brought to the author's attention as of version~4, report further improvements, including certificates above $\\delta>0.035$ and beyond $\\delta>0.036$. Some of these improvements may rely not only on larger prime ranges but also on modified constraint systems and additional degrees of freedom that deviate from Sawin's original formulation. Beyond this application, the work illustrates how randomized optimization heuristics can improve, verify, and refine explicit certificates for combinatorial geometry through nonlinear integer optimization.", "date": "2026-06-02", "categories": []} {"id": "2606.10321", "title": "Baseline-Free Policy Optimization for Neural Combinatorial Optimization", "abstract": "Neural combinatorial optimization (NCO) trains autoregressive policies to solve routing problems. The standard training algorithm, REINFORCE with a rollout baseline, requires maintaining and periodically updating a frozen copy of the policy for variance reduction. This baseline introduces a structural vulnerability: on harder instances, a poor baseline produces noisy gradient estimates that can destabilize training. We evaluate Group Relative Policy Optimization (GRPO), an algorithm from large language model alignment that eliminates the baseline entirely by normalizing advantages within groups of sampled trajectories. In a controlled comparison of five RL algorithms on TSP and CVRP benchmarks within the RL4CO framework, we find that: (i) GRPO avoids the training collapse observed with REINFORCE on TSP-100, where performance degrades from cost 9.8 to 52.1 immediately after the warmup phase and does not recover under extended training; (ii) at matched gradient updates, GRPO achieves solution quality within 2% of POMO, a strong AM-based multi-start baseline, while requiring no external baseline; and (iii) P3O, a pairwise preference algorithm also from the alignment literature, is competitive on TSP but shows higher variability on CVRP. These results identify GRPO as a promising baseline-free alternative for NCO, particularly in settings where baseline-dependent training becomes fragile.", "date": "2026-06-09", "categories": []} {"id": "2606.10431", "title": "Vision-Assisted Foundation Model for Solving Multi-Task Vehicle Routing Problems", "abstract": "Multi-task vehicle routing problems play a critical role in enhancing efficiency across various industries and service sectors. These problems consist of multiple variants that optimize routing costs while meeting diverse customer constraints. Existing multi-task VRP solvers solely utilize a graph-based modality, limiting their ability to address variants with multiple constraints. As a format to represent complex semantics, vision modality shows great potential for encoding diverse VRP constraints. This motivates us to learn patch-level semantics from the vision images, and then integrate them into a graph-based model to solve various VRP variants simultaneously. However, directly applying this approach to multi-task VRPs presents three challenges: 1) existing VRP images lack constraint representations, which are essential for multi-task VRPs, 2) the fixed receptive field of individual patches cannot effectively accommodate varying requirements across tasks, and 3) imbalanced pixel distribution among constraints may cause the model to overlook constraints with fewer pixels. In this paper, we propose a vision-assisted foundation model (VaFM) to address these challenges. In the vision modality, input images tailored to all constraints are encoded by a convolutional neural network. The obtained patch embeddings are fused with graph-based nodes to generate solutions, with an auxiliary task designed to address the pixel-imbalanced issue. The performance of VaFM is evaluated across 16 different VRP variants. The experimental results demonstrate the superiority of VaFM over state-of-the-art methods, especially for variants with complex constraints.", "date": "2026-06-09", "categories": []} {"id": "2606.10611", "title": "Geometry-Aware Reinforcement Learning for 2D Irregular Nesting", "abstract": "Traditional heuristic solvers for the 2D irregular nesting problem share a fundamental limitation: they are blind to polygon geometry, relying on guided brute-force to navigate the continuous placement space with minimal geometrical guidance. In this paper, we argue that Reinforcement Learning is uniquely positioned to overcome this bottleneck. By pairing an optimization policy with a geometry-aware neural encoder, an agent can automatically discover rich geometric priors directly from data, utilizing these learned intuitions to strategically guide exploration. To realize this, we introduce the Polygons Transformer (PoT), a novel architecture that encodes 2D continuous vector geometries while allowing cross-polygons attention. We couple this novel architecture with a Combinatorial Optimization Reinforcement Learning (CORL) training framework to find optimal solutions. To support this paradigm, we release an open-source training dataset derived from complex geographic contours alongside a dedicated evaluation benchmark. Our empirical validation demonstrates that our trained agent achieves area utilization performance highly competitive with Sparrow, the state-of-the-art heuristic solver, proving that reinforcement learning can successfully discover and exploit geometric awareness for precise spatial tasks.", "date": "2026-06-09", "categories": []} {"id": "2412.13858", "title": "IDEQ -- Improving Diffusion Models for the Traveling Salesman Problem (TSP) by Leveraging the Structure of the Solution Space", "abstract": "We investigate diffusion models to solve the Traveling Salesman Problem. Building on the recent DIFUSCO and T2TCO approaches, we propose IDEQ. IDEQ improves the quality of the solutions by leveraging the constrained structure of the state space of the TSP. Another key component of IDEQ consists in replacing the last stages of DIFUSCO curriculum learning by considering a uniform distribution over the Hamiltonian tours whose orbits by the 2-opt operator converge to the optimal solution as the training objective. Our experiments show that IDEQ improves the state of the art for such neural network based techniques on synthetic instances. More importantly, our experiments show that IDEQ performs very well on the instances of the TSPlib, a reference benchmark in the TSP community: it closely matches the performance of the best heuristics, LKH3, being even able to obtain better solutions than LKH3 on 2 instances of the TSPlib defined on 1577 and 3795 cities. IDEQ obtains 0.3% optimality gap on TSP instances made of 500 cities, and 0.5% on TSP instances with 1000 cities. This sets a new SOTA for neural based methods solving the TSP. Moreover, IDEQ exhibits a lower variance and better scales-up with the number of cities with regards to DIFUSCO and T2TCO.", "date": "2024-12-18", "categories": []} {"id": "2601.01665", "title": "Adversarial Instance Generation and Robust Training for Neural Combinatorial Optimization with Multiple Objectives", "abstract": "Deep reinforcement learning (DRL) has shown great promise in addressing multi-objective combinatorial optimization problems (MOCOPs). Nevertheless, the robustness of these learning-based solvers has remained insufficiently explored, especially across diverse and complex problem distributions. In this paper, we propose a unified robustness-oriented framework for preference-conditioned DRL solvers for MOCOPs. Within this framework, we develop a preference-based adversarial attack to generate hard instances that expose solver weaknesses, and quantify the attack impact by the resulting degradation on Pareto-front quality. We further introduce a defense strategy that integrates hardness-aware preference selection into adversarial training to reduce overfitting to restricted preference regions and improve out-of-distribution performance. The experimental results on multi-objective traveling salesman problem (MOTSP), multi-objective capacitated vehicle routing problem (MOCVRP), and multi-objective knapsack problem (MOKP) verify that our attack method successfully learns hard instances for different solvers. Furthermore, our defense method significantly strengthens the robustness and generalizability of neural solvers, delivering superior performance on hard or out-of-distribution instances.", "date": "2026-01-04", "categories": []} {"id": "2601.06188", "title": "Dynamic Distributed Constraint Optimization and Metareasoning for Continual, Large-Scale Satellite Operations", "abstract": "As Earth-observing satellite constellations grow in size and capability, distributed onboard control offers a pathway to novel responses and time-sensitive measurements. However, deploying autonomy to satellites requires efficient computation and communication. This work addresses the challenge of scheduling observations for hundreds of satellites in a dynamic, large-scale problem with millions of variables. We present the dynamic multi-satellite constellation observation scheduling problem (DCOSP), a new formulation of dynamic distributed constraint optimization problems (DDCOP) that models integrated scheduling and execution. DCOSP features a novel optimality condition, for which we construct an exact omniscient offline algorithm. Motivated by the strong resource constraints of onboard satellite operations, we introduce a framework to incorporate metareasoning in DDCOPs that controls when agents expend resources to recompute solutions. In addition, we present the dynamic incremental neighborhood stochastic search (D-NSS) algorithm, an incomplete online decomposition-based DDCOP algorithm that repairs localized sub-problems in response to dynamic events. We demonstrate in realistic simulations that D-NSS converges to near-optimal solutions, outperforming standard DDCOP baselines in solution quality, computation time, and message volume, while our metareasoning framework successfully balances resource conservation with utility. As part of the NASA FAME mission, this work lays the foundation for the largest in-space demonstration of distributed multi-agent AI to date.", "date": "2026-01-08", "categories": []} {"id": "2604.23053", "title": "ML-Guided Primal Heuristics for Mixed Binary Quadratic Programs", "abstract": "Mixed Binary Quadratic Programs (MBQPs) are an important and complex set of problems in combinatorial optimization. As solving large-scale combinatorial optimization problems is challenging, primal heuristics have been developed to quickly identify high-quality solutions within a short amount of time. Recently, a growing body of research has also used machine learning to accelerate solution methods for challenging combinatorial optimization problems. Despite the increasing popularity of these ML-guided methods, a large body of work has focused on Mixed-Integer Linear Programs (MILPs). MBQPs are challenging to solve due to the combinatorial complexity coupled with nonlinearities. This work proposes ML-guided primal heuristics for Mixed Binary Quadratic Programs (MBQPs) by adapting and extending existing work on ML-guided MILP solution prediction to MBQPs. We introduce a new neural network architecture for MBQP solution prediction and a new training data collection procedure. Moreover, we extend existing loss functions in solution prediction and propose to combine contrastive and weighted cross-entropy losses. We evaluate the methods on standard and real-world MBQP benchmarks and show that the developed ML-guided methods significantly outperform existing primal heuristics and state-of-the-art solvers. Furthermore, models trained with our proposed extension with combined losses outperform other ML-based methods adapted from MILPs and improve generalization in cross-regional inference on a real-world wind farm layout optimization problem.", "date": "2026-04-24", "categories": []} {"id": "2606.09343", "title": "Leveraging Structural Constraints for Diffusion-based Neural TSP Solvers", "abstract": "Neural combinatorial optimization has recently achieved strong results on the Euclidean Traveling Salesman Problem (TSP) using generative models such as diffusion and consistency models. State-ofthe-art approaches like FT2T combine fast consistency-based prediction with gradient-based inference time refinement. However, gradient search often incurs significant computational overhead and may not align with the discrete structure of feasible solutions. We introduce Projected Consistency Inference (PCI), a plug-and-play, retraining-free alternative that replaces gradient refinement with structure-aware projections: PCI decodes valid Hamiltonian tours from the consistency model output and applies a lightweight local search (e.g., 2-opt). PCI achieves an average optimality gap (OG) of 0.17% on TSP with 500 cities, and 0.31% on TSP with 1000 cities, outperforming FT2T best settings (OG 0.22% and 0.36%, respectively) while reducing the inference time up to 30 to 40%. PCI also exhibits lower variance and memory usage, and can surpass classical heuristics such as LKH3 in rapid solution generation. Our results demonstrate that structure-aware inference time operations provide a practical and principled path for neural TSP solvers, complementing training time objectives.", "date": "2026-06-08", "categories": []} {"id": "2606.09623", "title": "Constrained user-item allocation for e-commerce marketing campaigns", "abstract": "When running marketing campaigns, retailers must decide which products to promote and which users to target. These decisions are inherently coupled: effective campaigns match users and items with strong mutual affinity into non-overlapping groups of predefined sizes. However, existing approaches assume predefined campaign structure or decouple item selection from user assignment, and cannot discover campaign groupings directly from joint interaction patterns. We therefore formalize this campaign problem as auto-targeting: jointly selecting users and items to construct multiple disjoint campaigns. To solve this combinatorial problem, we propose three complementary strategies: (i) constrained spectral biclustering to find dense regions in the user-item affinity matrix, (ii) greedy local search with pairwise swaps for combinatorial refinement, and (iii) a multi-armed bandit framework to escape local optima through exploration. We evaluate these methods on a synthetic dataset, the Amazon Reviews benchmarks, and large-scale proprietary commercial data, and compare the results to simulated annealing as a baseline. The results show that biclustering consistently achieves the highest campaign quality, lift, and fairness scores. While biclustering runs efficiently on smaller datasets, its runtime increases substantially on very large ones, where bandit-based methods instead offer a scalable alternative.", "date": "2026-06-08", "categories": []} {"id": "2605.31051", "title": "Linear Ordering Problem: Time for a Change", "abstract": "The Linear Ordering Problem (LOP) is a fundamental combinatorial optimization problem with important applications in areas such as economics, social choice, and machine learning. Its most prominent use is the triangulation of economic input-output tables, which helps identify critical industries in an economy. Most existing algorithms have been evaluated on benchmarks derived from outdated macroeconomic data, which no longer reflect the structure of contemporary economies. Furthermore, LOP instances often exhibit many distinct global optima that can differ substantially from one another, creating challenges for applications that rely on a single solution. To address these limitations, we introduce a novel benchmark suite derived from up-to-date real-world economic data and an algorithmic scheme that leverages state-of-the-art LOP metaheuristics to generate diverse sets of high-quality solutions, together with metrics for assessing both quality and diversity. Experiments were conducted to report results on the proposed benchmark suite under both the traditional single-solution setting and the newly introduced multi-solution scenario", "date": "2026-05-29", "categories": []} {"id": "2606.06518", "title": "DiBS: Diffusion-Informed Branch Selection", "abstract": "Sudoku is a representative constraint satisfaction problem that requires global structural reasoning under strict discrete constraints. The existing works of solving Sudoku mainly focus on two dominant approaches, i.e., traditional heuristic and deep learning solver. However, they suffer from two complementary limitations: learning-based solvers lack hard correctness guarantees, while complete symbolic solvers are still prone to long-tail search. To address these shortcomings, we propose a novel diffusion model-guided approach, termed as DiBS, for the branch selection search process. Specifically, DiBS keeps the symbolic solver complete and uses the diffusion model as a branch-ordering guide. The core method is ranking candidate values under the current partial assignment and lightweight consistency signal. Furthermore, we provide an in-depth theoretical proof to reveal how it works and why it works. Experiments on the challenging Royle 17-clue Sudoku benchmark show that our DiBS substantially reduces search cost relative to strong heuristic baselines, especially in nodes, backtracks, and long-tail percentiles. Besides, these results confirm that learned global guidance is effective on hard instances where branch-order mistakes are most expensive. All codes are available at https://github.com/shanxierdan/DiBS.", "date": "2026-06-02", "categories": []} {"id": "2606.06543", "title": "Coordinated optimization of departure sequencing and section-track allocation in railway short-term concentrated departure scenarios based on qubo and hybrid quantum algorithms", "abstract": "This study examines the coordinated optimization of departure sequencing and section-track allocation in railway short-term concentrated departure scenarios. A quadratic unconstrained binary optimization (QUBO) model is formulated to represent departure-position assignment and section-track selection within a unified binary framework. Because the quality of a dispatching scheme depends on time-dependent operational interactions that cannot be fully captured by a static combinatorial model, a simulation-based evaluation layer is introduced to assess section occupation, intermediate-station waiting, platform-capacity pressure, running-time fluctuations, and delay propagation. Within this layered framework, conventional heuristics, quantum-inspired algorithms, and hybrid algorithms are compared on the same decision structure. The results show that the QUBO model can generate feasible candidate schemes after decoding, while the simulation layer clearly differentiates the operational performance of the competing algorithms under both normal and disturbed conditions. In the tested scenarios, QPSO-QAOA performs best under normal conditions, and the quantum-enhanced methods reduce comprehensive cost by 4.28\\%--26.26\\% and total delay by 4.37\\%--24.25\\% on average under dynamic conditions relative to their conventional counterparts. These findings suggest that the integration of QUBO-based modeling and simulation-based evaluation provides a useful methodological framework for railway short-term concentrated departure scheduling, although validation with real operational data remains necessary.", "date": "2026-06-04", "categories": []} {"id": "2606.07403", "title": "The Proxy Benders Decomposition", "abstract": "Benders decomposition is a fundamental framework for solving large-scale mixed-integer optimization problems with complicating variables that, when fixed, yield significantly easier subproblems. However, classical Benders decomposition repeatedly solves highly similar subproblems and often exhibits zigzagging behavior across iterations, leading to slow convergence in large-scale settings. Motivated by the repetitive structure and parametric nature of Benders subproblems, this paper introduces the proxy Benders decomposition (Proxy-BD), a new decomposition framework in which subproblem optimization is replaced by certified optimization proxies rather than repeated exact solves.\n The proposed proxy follows a self-supervised predict-project-and-complete mechanism that produces dual-feasible solutions for generating provably valid Benders cuts. The framework preserves the theoretical validity of the decomposition independently of prediction quality through a projection-and-completion certification layer. A formal characterization of proxy-induced cuts is established, and the framework naturally extends to modern decomposition schemes, including branch-and-Benders-cut algorithms.\n Computational experiments on large-scale facility location and network design problems demonstrate that Proxy-BD substantially reduces the computational effort of subproblems while maintaining near-optimal solution quality. On large-scale uncapacitated facility location instances up to 2000x2000, Proxy-BD achieves median optimality gaps below 0.5%, yields up to 161x median speedups, and reduces the number of generated cuts by more than 240x on the largest instances. The computational gains consistently increase with recourse complexity, indicating that proxy-based inference scales substantially more favorably than repeated exact subproblem optimization in large-scale decomposition settings.", "date": "2026-06-05", "categories": []} {"id": "2605.29916", "title": "Selection Hyper-heuristics Can Automatically Adjust the Learning Period to Optimally Solve Pseudo-Boolean Problems", "abstract": "The Random Gradient hyper-heuristic was recently shown to be able to learn the optimal neighbourhood size when optimizing the LeadingOnes benchmark via the Randomised Local Search (RLS) meta-heuristic. However, for this to happen, a learning period of a certain length $\\tau$ had to be used, differently from classic hyper-heuristics, which change their behaviour based on the success of only the previous iteration. In this paper, we show how to automatically set this new parameter value, relieving the user from the non-trivial task of controlling this novel algorithm parameter. We prove that the resulting hyper-heuristic selects the optimal neighbourhood size in a $1-o(1)$ fraction of the iterations and, consequently, optimises the LeadingOnes benchmark in the best possible time (apart from lower-order terms) achievable with these neighborhood sizes.", "date": "2026-05-28", "categories": []} {"id": "2606.05956", "title": "Bidirectional Search for Longest Paths: Case for Front-to-Front Heuristics", "abstract": "Bidirectional heuristic search can potentially reduce search effort for problems amenable to backward search. Therein, it is well-known that front-to-front heuristics can reduce the number of node expansions, but their overhead is so high that overall runtime almost always increases. We propose BiXDFBnB, a bidirectional depth-first branch-and-bound algorithm that adapts the Single-Frontier Bidirectional Search (SFBDS) framework - originally developed for shortest-path (MIN) problems - to the Generalized Longest Simple Path (GLSP) setting. Because SFBDS inherently operates on paired states, front-to-front (F2F) heuristic evaluation arises naturally and avoids the overhead typically associated with bidirectional frontier management. We show that this adaptation can be successfully applied to maximization (MAX) problems while efficiently handling overlapping constraints. BiXDFBnB is applied to several types of longest-path problems: Longest Simple Path (LSP), Snakes, and Coil-in-the-Box (CIB). Empirical evaluation shows that the new algorithm frequently reduces the number of node expansions and, in some cases, also improves overall runtime.", "date": "2026-06-04", "categories": []} {"id": "2606.04039", "title": "Beyond Static Priors: Dynamic Neural Guidance for Large-Scale Ant Colony Optimization", "abstract": "Neural-guided Ant Colony Optimization (ACO) suffers from a fundamental training-inference misalignment: policies are typically trained to generate static priors (e.g., heatmaps), yet deployed to guide iterative, long-horizon search processes. In this paper, we present DyNACO, a novel framework that achieves dynamic neural guidance by periodically observing the pheromone distribution and the incumbent solution. To make DyNACO tractable at scale, we pair the policy with a perturbation-based ACO backend and a scope-restricted refinement mechanism that jointly ensure efficacy and stable credit assignment. On TSP, DyNACO scales to 100,000-node instances and outperforms neural baselines while often reducing total runtime compared to the unguided solver. We extend DyNACO to CVRP via a capacity-aware backend, consistently improving the unguided baseline with less than 1% neural overhead. We further provide in-depth analysis validating the model's generalization capabilities and elucidating why dynamic guidance outperforms static priors. Our work underscores the necessity of aligning neural training with iterative search dynamics in learning-guided optimization. The code is available at https://github.com/shoraaa/DyNACO.", "date": "2026-06-02", "categories": []} {"id": "2606.04860", "title": "Learning Empirically Admissible Neural Heuristics for Combinatorial Search", "abstract": "Finding optimal solution paths for combinatorial puzzles like the Rubik's Cube, sliding tile puzzles, and Lights Out remains a classical challenge in artificial intelligence. Heuristic search algorithms, such as A* , guarantee path optimality only when using an admissible heuristic-one that never overestimates the true remaining cost-to-go. Deep reinforcement learning (RL) methods like DeepCubeA train deep neural networks to approximate cost-to-go heuristics. However, standard mean-squared error (MSE) training regularly yields overestimations, violating admissibility and compromising solution optimality. In this paper, we introduce a generalizable framework for learning validation-calibrated admissible neural heuristics. We train a value network using an underestimating Admissible Bellman Operator combined with an Asymmetric Loss function to penalize overestimation. To account for residual neural function approximation errors, we propose a post-hoc calibration safety offset computed over validation scrambles. We demonstrate that our calibrated neural heuristics achieve no observed admissibility violations under the evaluation protocol and preserve path optimality in practice while reducing search node expansions by up to 83.0% on a 2 by 2 Rubik's Cube, 19.9% on a 3 by 3 Lights Out grid, and 1.9% on an 8-Puzzle compared to standard analytical baselines.", "date": "2026-06-03", "categories": []} {"id": "2606.04946", "title": "A General Framework for Dynamic Consistent Submodular Maximization", "abstract": "Consistency is an important property in dynamic submodular maximization and entails maintaining a near-optimal solution at all times, making only a small number of adjustments to the solution in each step. Prior work has explored this question for the insertion-only case, where the algorithm faces a stream of $n$ insertions, and has established lower and upper bounds for the cardinality-constrained version of the problem. We consider this question in the fully dynamic setting, where the stream of operations may contain both insertions and deletions.\n We develop a general framework for designing algorithms for this setting, and instantiate it to obtain the first constant-factor approximations with sublinear consistency. For cardinality constraints, we propose a $\\frac 12 - O(\\varepsilon)$ approximation that is $O\\left(\\frac{1}{\\varepsilon^2}\\right)$ consistent. For rank-$k$ matroid constraints, we construct a $\\frac 14 - O(\\varepsilon)$ approximation to the dynamic optimum that is $O\\left(\\frac{\\log k}{\\varepsilon^2}\\right)$ consistent.", "date": "2026-06-03", "categories": []} {"id": "2501.17377", "title": "ASAP: Exploiting the Satisficing Generalization Edge in Neural Combinatorial Optimization", "abstract": "Deep Reinforcement Learning (DRL) has emerged as a promising approach for solving Combinatorial Optimization (CO) problems, such as the 3D Bin Packing Problem (3D-BPP), Traveling Salesman Problem (TSP), or Vehicle Routing Problem (VRP), but these neural solvers often exhibit brittleness when facing distribution shifts. To address this issue, we uncover the Satisficing Generalization Edge, which we validate both theoretically and experimentally: identifying a set of promising actions is inherently more generalizable than selecting the single optimal action. To exploit this property, we propose Adaptive Selection After Proposal (ASAP), a generic framework that decomposes the decision-making process into two distinct phases: a proposal policy that acts as a robust filter, and a selection policy as an adaptable decision maker. This architecture enables a highly effective online adaptation strategy where the selection policy can be rapidly fine-tuned on a new distribution. Concretely, we introduce a two-phase training framework enhanced by Model-Agnostic Meta-Learning (MAML) to prime the model for fast adaptation. Extensive experiments on 3D-BPP, TSP, and CVRP demonstrate that ASAP improves the generalization capability of state-of-the-art baselines and achieves superior online adaptation on out-of-distribution instances.", "date": "2025-01-29", "categories": []} {"id": "2503.03137", "title": "Learning to Reduce Search Space for Generalizable Neural Routing Solver", "abstract": "Constructive neural combinatorial optimization (NCO) offers a promising paradigm for solving vehicle routing problems (VRPs) by directly learning to construct approximate optimal solutions, thereby reducing reliance on expert knowledge for algorithm design. However, scaling these methods to handle large-scale instances remains challenging due to high computational complexity. While recent dynamic search space reduction (SSR) methods can improve inference efficiency through geometric distance-based pruning, they often struggle on complex instances with non-uniform distributions or when optimal solutions rely heavily on non-spatial constraints. To address this critical issue, we propose Learning to Reduce (L2R), which is the first learning-based dynamic SSR framework. L2R learns to adaptively prioritize nodes by extracting patterns from problem-specific features to prune the search space at each step, enabling efficient and scalable solution construction. Extensive experiments show that our L2R framework generalizes robustly to different problem scales and data distributions on various VRP variants. To the best of our knowledge, L2R is the first neural solver to effectively scale to VRP instances with $10$ million nodes while maintaining high solution quality, which significantly pushes the frontier of NCO in terms of generalization and scalability. Our code is available at https://github.com/CIAM-Group/L2R.", "date": "2025-03-05", "categories": []} {"id": "2510.03086", "title": "Chaining 2-FWL GNNs for Combinatorial Graph Alignment", "abstract": "For the combinatorial graph alignment problem (GAP) -- finding the node correspondence that maximizes the number of common edges (nce) between two unlabeled graphs -- properly initialized FAQ remains a strong classical baseline, while existing GNN approaches struggle in the purely structural setting. We introduce a chaining procedure: a sequence of Folklore-type (2-FWL) GNNs in which each network is trained with cross-entropy after decoding the previous network's similarity matrix and ranking nodes by their current alignment quality. This non-differentiable ranking step injects discrete combinatorial feedback at every link; at inference, we iterate the final network and keep the candidate with highest observed nce. On sparse Erdos-Renyi graphs at noise level 0.25, chained FGNNs with FAQ post-processing reach 85% accuracy versus 13% for FAQ initialized from the convex relaxation, and essentially 0% for prior GNN methods. On correlated regular graphs, where MPNNs with constant features produce identical node embeddings (1-WL fails to refine) and FAQ's convex initialization is degenerate, chaining is the only method we know that recovers a non-trivial alignment. On three real-world benchmarks (yeast PPI, coauthorship, and road networks), we show that recent comparisons underestimate FAQ by initializing it from a uniform doubly stochastic matrix; once FAQ is initialized from the convex relaxation it already surpasses prior reported numbers, and dataset-specific chained FGNNs further improve on this strengthened baseline.", "date": "2025-10-03", "categories": []} {"id": "2602.08585", "title": "Predicting Future Utility: Global Combinatorial Optimization for Task-Agnostic KV Cache Eviction", "abstract": "Given the quadratic complexity of attention, KV cache eviction is vital to accelerate model inference. Current KV cache eviction methods typically rely on instantaneous heuristic metrics, implicitly assuming that score magnitudes are consistent proxies for importance across all heads. However, this overlooks the heterogeneity in predictive fidelity across attention heads. While certain heads prioritize the instantaneous contribution of tokens, others are dedicated to capturing long-horizon utility. In this paper, we propose that optimal budget allocation should be governed by the marginal utility in preserving long-term semantic information. Building on this insight, we propose LU-KV, a novel framework that formulates head-level budget allocation as a global combinatorial optimization problem to maximize the long-horizon marginal contribution of reserved tokens. To solve this non-convex problem, we employ a convex-hull relaxation and a marginal-utility-based greedy solver, achieving near-optimal solutions. Furthermore, we implement a data-driven offline profiling protocol to facilitate the practical deployment of LU-KV. Evaluations on LongBench and RULER benchmarks demonstrate that LU-KV reduces KV cache size by 80% with minimal performance degradation, while also decreasing inference latency and GPU memory footprint.", "date": "2026-02-09", "categories": []} {"id": "2605.09382", "title": "Learning-Augmented Scalable Linear Assignment Problem Optimization via Neural Dual Warm-Starts", "abstract": "The Linear Assignment Problem is a fundamental combinatorial optimization task where classical exact solvers ensure optimality but suffer from an $\\mathcal{O}(N^{3})$ bottleneck, while recent neural approximations struggle with scalability and exactness. We propose a learning-augmented framework that accelerates exact solvers by predicting dual variables to warm-start the search, backed by a fallback mechanism to preserve worst-case guarantees. Central to our approach is RowDualNet, a lightweight, row-independent architecture that avoids the $\\mathcal{O}(N^{2})$ memory bottleneck of graph models, enabling scalable neural warm-starting up to $N=16{,}384$. Feasibility is guaranteed by construction via the Min-Trick mechanism, completely eliminating the need for costly iterative projections. Empirically, our method drastically reduces the search effort of the Jonker-Volgenant (LAPJV) algorithm, yielding robust zero-shot generalization with strict optimality and end-to-end speedups of over 2x on complex synthetic data, 1.25x on real-world tracking, and 1.5x on transportation networks.", "date": "2026-05-10", "categories": []} {"id": "2606.00718", "title": "LLM-Driven Co-Evolutionary Automated Heuristic Design for Bi-Component Coupled Combinatorial Optimization", "abstract": "While Large Language Models (LLMs) have recently shown promise in Automated Heuristic Design (AHD), existing methods typically generate and evolve heuristics as a single operator or search strategy, limiting their ability to model strong coupling among multiple decision substructures in problems such as the Traveling Thief Problem (TTP) and the Traveling Purchaser Problem (TPP). In this work, we propose CoEvo-AHD, an LLM-driven dual-population co-evolutionary framework for automated heuristic design in coupled combinatorial optimization. Unlike prior methods that evolve individual heuristics in isolation, CoEvo-AHD leverages LLMs to co-evolve two closely related operator populations. A cooperative evaluation mechanism explicitly captures interactions between route and selection operators, while pairwise scoring and synergistic joint crossover help discover complementary operator logic for joint improvement across coupled decision subspaces. We further design a tool-invocation environment library that encapsulates frequently used core operations, such as local-search delta computation, into callable functions, enabling LLM-generated operators to use standardized interfaces instead of reimplementing inefficient and error-prone problem-specific loops. Experiments on TTP and TPP show that CoEvo-AHD automatically discovers cooperative heuristic combinations and achieves competitive solution quality against traditional heuristics.", "date": "2026-05-30", "categories": []} {"id": "2606.01084", "title": "MViewRouter: Internalizing Geometric Equivariance via Multi-view Alternating Attention for Combinatorial Routing", "abstract": "Combinatorial routing problems such as the Traveling Salesman Problem (TSP) and the Capacitated Vehicle Routing Problem (CVRP) are fundamental NP-hard problems with broad real-world applications. While recent deep reinforcement learning methods have shown promising performance, they typically handle geometric symmetries only through data augmentation, resulting in inconsistent decisions and limited generalization. To address this issue, we propose MViewRouter, a multi-view framework that internalizes geometric equivariance as a structural inductive bias to achieve invariant decision-making across routing problem variants. Our approach introduces a Multi-view Alternating Attention (MAA) mechanism that enables parallel processing over the $D_4$ symmetry group, alternating between intra-view relational modeling and inter-view feature alignment. Furthermore, we optimize the policy via Collective Policy Gradient Aggregation (CPGA), leveraging consensus gradients from multiple symmetric views to stabilize training and accelerate convergence. Experiments on TSP and CVRP benchmarks, as well as real-world TSPLIB instances, demonstrate that MViewRouter achieves competitive solution quality and strong zero-shot generalization.", "date": "2026-05-31", "categories": []} {"id": "2606.01425", "title": "Learning-based Directed Graph Abstraction of Combinatorial Spaces for Order-Preserving Search in Mixed-Combinatorial Nonlinear Optimization", "abstract": "Mixed-combinatorial nonlinear programming (MCNLP) problems arise in many engineering design and planning applications, e.g., due to categorical, component, and geometric design choices, as well as joint task and motion planning. Traditional representations of combinatorial spaces, such as integer or binary encoding, often introduce spurious relations, increase dimensionality, and require additional compatibility constraints. Instead, this paper draws on recent developments in robot planning and vehicle/network routing domains that aim to learn search heuristics over combinatorial spaces using graph neural networks (GNNs). More specifically, this paper presents a first-of-its-kind structured abstraction of the combinatorial space by learning a mapping from an undirected fully connected graph of combinations to a directed graph indicating improvement directions using an Edge Field Graph Network (EFGN). To demonstrate the utility of this new way of abstracting the combinatorial space in solving MCNLPs, we adopt a recent optimization framework that purely searches over the non-combinatorial (e.g., continuous) variables and retrieves the best-suited combination for each candidate design by using the abstraction model, akin to a recommender system. The presented direction-aware abstraction model provides a potentially more scalable and interpretable retrieval of combinations compared to the original recommendation system in that framework. For evaluation, the proposed method is integrated with a well-known particle swarm optimization and genetic algorithm solvers on three benchmark nonlinear problems with varying numbers of combinations and variables. Compared to baseline solvers using indexified combinations, the GNN-based recommender consistently achieves better mean optimum values and robustness across multiple runs.", "date": "2026-05-31", "categories": []} {"id": "2606.01987", "title": "Graph Edit Distance Formulation for the Vehicle Routing Problem: Theory and Analysis", "abstract": "We show that the Vehicle Routing Problem (VRP) can be reformulated as a Graph Edit Distance (GED) maximization problem. Under a simple edge-deletion cost model, minimizing total route cost is equivalent to maximizing the total weight of edges deleted from the complete instance graph. This formulation models VRP at the edge level, where solutions are defined by selected edges rather than route sequences, enabling structural analyses that are difficult in classical formulations: per-edge attribution of solution quality, decomposition of the optimality gap, characterization of solution sparsity, and identification of edges that are hard to reach by greedy construction. Theoretically, we establish a merge-decomposition theorem showing that Clarke-Wright savings equal per-merge GED increments, and an approximation-transfer theorem that turns GED approximation ratios into VRP cost bounds. Using this reformulation, we analyze 90 CVRP benchmark instances with known optimal solutions. We find that optimal routing graphs use only 5.5% of available edges, that approximately 3.0% of optimal edges are consistently not found by Clarke-Wright heuristics under repeated restarts, and that the cost gap decomposes into missed optimal edges and substituted non-optimal edges of comparable total weight. The edge-additive objective provides a natural per-edge supervision signal for future graph neural network approaches to edge prediction, suggesting a potential connection to graph neural network approaches that we leave for follow-up work.", "date": "2026-06-01", "categories": []} {"id": "2606.02136", "title": "Edge-aware Decoding for Neural Asymmetric Routing", "abstract": "Neural asymmetric routing models increasingly encode directionality through matrix representations and asymmetry-aware attention. The final routing action, however, is not a node in isolation but a directed transition chosen under the current partial route. This creates a representation--decision mismatch: pairwise cost information may be encoded upstream while the final candidate logit is still largely parameterized as context--node compatibility. We propose a decoder-design principle for neural asymmetric routing: the final score should explicitly expose transition-level quantities suggested by the problem's cost-to-go structure. We instantiate this principle with an edge-aware decoder that adds candidate-specific terms for the current directed edge, return-to-start closure, and static lightweight lookahead, while keeping the representation backbone fixed. On a controlled SVD/Sinkhorn asymmetric backbone, the decoder improves over the RADAR reference when trained on ATSP-100 and evaluated zero-shot on ATSP-100/200/500/1000, reducing the ATSP-1000 gap from $4.13\\%$ to $2.73\\%$. On ACVRP, the same score-level modification shows the same qualitative trend under a richer routing state. ATSP ablations and directed-transition diagnostics sharpen the mechanism: the strongest evidence concerns sensitivity to the current directed edge, while closure and static lookahead act as heuristic continuation cues. The results support a mechanism study: a key decoder-side signal in neural asymmetric routing is decision-time exposure of transition-level edge information.", "date": "2026-06-01", "categories": []} {"id": "2606.02294", "title": "Regularized Large Neighborhood Search", "abstract": "Operations research practitioners typically tackle NP-hard combinatorial problems using large neighborhood search (LNS), a scalable heuristic that iteratively refines a current solution by locally re-optimizing subsets of its variables. In contrast, most existing approaches for integrating combinatorial optimization layers into neural networks still assume access to an exact global solution, which is computationally intractable. We bridge this gap by introducing regularized LNS (RLNS). By regularizing or perturbing local subproblems, we turn the LNS heuristic into an efficient MCMC sampler over the combinatorial set of feasible solutions, with associated Fenchel-Young losses. Under entropic regularization, we prove that RLNS performs exact block Gibbs sampling. Furthermore, adjusting the number of RLNS iterations allows us to interpolate between pseudolikelihood and exact maximum likelihood estimation, for end-to-end learning without global solvers. We demonstrate our approach on $k$-subset selection, generalized assignment, and stochastic vehicle scheduling problems.", "date": "2026-06-01", "categories": []} {"id": "2310.14161", "title": "Promoting Generalization for Exact Solvers via Adversarial Instance Augmentation", "abstract": "Machine learning has been successfully applied to improve the efficiency of Mixed-Integer Linear Programming (MILP) solvers. However, the learning-based solvers often suffer from severe performance degradation on unseen MILP instances -- especially on large-scale instances from a perturbed environment -- due to the limited diversity of training distributions. To tackle this problem, we propose a novel approach, which is called Adversarial Instance Augmentation and does not require to know the problem type for new instance generation, to promote data diversity for learning-based branching modules in the branch-and-bound (B&B) Solvers (AdaSolver). We use the bipartite graph representations for MILP instances and obtain various perturbed instances to regularize the solver by augmenting the graph structures with a learned augmentation policy. The major technical contribution of AdaSolver is that we formulate the non-differentiable instance augmentation as a contextual bandit problem and adversarially train the learning-based solver and augmentation policy, enabling efficient gradient-based training of the augmentation policy. To the best of our knowledge, AdaSolver is the first general and effective framework for understanding and improving the generalization of both imitation-learning-based (IL-based) and reinforcement-learning-based (RL-based) B&B solvers. Extensive experiments demonstrate that by producing various augmented instances, AdaSolver leads to a remarkable efficiency improvement across various distributions.", "date": "2023-10-22", "categories": []} {"id": "2605.29005", "title": "LoRe: Adaptive Interaction-Evaluation Routing with Per-Step Interaction Budgets for Iterative Graph Solvers", "abstract": "Diffusion-based neural solvers for combinatorial optimization repeatedly re-evaluate dense edge/factor interactions, making inference expensive in wall-clock time and often memory-bound at scale. Inspired by the computational methodologies of many-body physics, we introduce LoRe, a training-free, inference-time drop-in wrapper that enforces per-step interaction-evaluation budgeting: at each iteration, it evaluates only a fixed fraction of interactions by dynamically routing computation to high-conflict or high-uncertainty interactions, instead of using a fixed sparsification (e.g., static kNN graphs or static masks). Under fully inclusive end-to-end wall-clock accounting, LoRe substantially improves scalability on the Maximum Independent Set (MIS) problem, extending feasible inference more than $3\\times$ beyond the baseline's out-of-memory limit, delivering a $\\sim 8\\times$ speedup and a $\\sim 12\\times$ peak-memory reduction, with solution quality preserved in this regime. Demonstrating cross-task generality on the large-scale Traveling Salesperson Problem (TSP) and zero-shot robustness to topology shifts, LoRe achieves a $\\sim 15\\times$ speedup at $n=1000$ with a $44\\times$ memory reduction and competitive tour quality.", "date": "2026-05-27", "categories": []} {"id": "2605.29262", "title": "Harmonizing Real-Time Constraints and Long-Horizon Reasoning: An Asynchronous Agentic Framework for Dynamic Scheduling", "abstract": "The Dynamic Flexible Job Shop Scheduling Problem (DFJSP) necessitates a trade-off between instant reaction to stochastic disturbances and global optimization of production goals. Conventional priority rules are insufficiently flexible to handle complex disruptions, whereas learning-based approaches often compromise interpretability or fail to generalize across problem scales. Although Large Language Models (LLMs) offer advanced reasoning capabilities to bridge this gap, their substantial inference latency is incompatible with the millisecond-level decision cycles of industrial control systems. To resolve this conflict, we introduce RACE-Sched, an asynchronous agent-based framework that decouples policy execution from logical reasoning via a dual-stream architecture. The Reactive Stream executes low-latency symbolic heuristics to enable real-time dispatching, while the parallel Deliberative Stream leverages an LLM to synthesize, validate, and evolve these rules. Candidate rules undergo rigorous testing in a sandbox and are deployed via atomic updates, ensuring safety without blocking the control loop. Additionally, a semantic rule repository indexes validated heuristics for retrieval-based initialization which enhances transferability across problem scales. Extensive evaluations on GEN-Bench, MK-Bench, and JMS-Bench demonstrate that RACE-Sched outperforms leading Deep Reinforcement Learning and other LLM-based baselines. This approach harmonizes real-time constraints with long-horizon reasoning to achieve superior solution quality and robust adaptation to dynamic events.", "date": "2026-05-28", "categories": []} {"id": "2605.29366", "title": "Solving Integer Linear Programming with Parallel Tempering", "abstract": "Integer Linear Programming (ILP) serves as a versatile framework for modeling a wide range of combinatorial optimization problems, typically addressed by sophisticated exact solvers or heuristics. While learning-based approaches have recently shown their effectiveness, they suffer from poor generalization to out-of-distribution instances and inherent dependence on external solvers. In this work, we propose a solver-free, sampling-based optimization framework for ILP that directly explores discrete feasible regions without training or external solvers. Exploiting the linear structure of ILP, we employ a Locally-Balanced Proposal to construct a transition kernel, thereby avoiding the gradient approximation. To overcome the highly multimodal nature of ILP energy landscapes, we integrate Parallel Tempering. In addition to standard temperature tempering, we introduce penalty tempering, which modulates constraint barriers while preserving the objective landscape over feasible solutions. Empirically, our method consistently outperforms SCIP across all four benchmarks, matches or exceeds Gurobi on two of four tasks within a 200-second budget, and is substantially more robust to distribution shift than learning-based methods. Furthermore, on MIPLIB 2017 instances, our framework remains competitive with classical solvers without any problem-specific tuning.", "date": "2026-05-28", "categories": []} {"id": "2605.30220", "title": "TriSearch: Learning to Optimize Triangulations via Bistellar Flips", "abstract": "We introduce TriSearch, a reinforcement learning framework for optimizing objectives over triangulations of a polytope via bistellar flips. The key idea is a circuit-supported subtriangulation action representation: feasible flips are encoded by their supporting circuit and realized local subtriangulation, enabling a learned policy to rank them using local geometric and combinatorial features. This yields a dimension-agnostic interface and enables efficient traversal of the flip graph without explicit enumeration of the full triangulation space. Instantiated in 3D and 4D, TriSearch generalizes zero-shot from small training instances to larger polytopes with exponentially larger search spaces. It achieves top performance on metric objectives in 3D and, in 4D, discovers more distinct Fine, Regular, Star triangulations of reflexive polytopes, corresponding to Calabi-Yau threefolds, than existing samplers under a fixed budget.", "date": "2026-05-28", "categories": []} {"id": "2605.27566", "title": "DynaSchedBench: Calibrated Dynamic Scheduling Benchmarks and Observability Paradox in LLM-based Scheduling Agents", "abstract": "Progress in neural combinatorial optimization for Dynamic Flexible Job Shop Scheduling Problem (DFJSP) is currently hindered by a methodological tension: static benchmarks encourage benchmark overfitting, while uncalibrated generators obscure algorithmic capability with stochastic noise. To resolve this, we introduce \\textbf{DynaSchedBench}, a diagnostic framework for DFJSP that rigorously controls the instance-generation process. Instead of relying on parameter sampling, our approach utilizes Sequential Event-Space Calibrator (SESC) that computes a novel Schedule Stress Index (SSI) to stratify instances by difficulty. We demonstrate that SESC is substantially more computationally efficient than evolutionary baselines while converging reliably to the target metrics. The framework integrates modular components for instance generation, snapshot-based simulation, agents, evaluation, and visualization, thereby enabling rigorous testing of reactive and lookahead-based policies. Leveraging this calibrated environment, we identify key limitations of LLM-based scheduling agents. Specifically, in step-wise online decision-making for dynamic scheduling, we identify an ``Observability Paradox'': providing agents with oracle access to full structural information can degrade policy performance, underperforming concise information. Furthermore, despite substantial token overhead, tool-augmented and refinement strategies fail to reliably improve performance, and most LLM agents fail to consistently surpass strong dispatching baselines-behaving more like robust heuristic approximators than superior optimizers.", "date": "2026-05-26", "categories": []} {"id": "2605.28337", "title": "An Enhanced Large Neighborhood Search Approach for the Capacitated Facility Location Problem with Incompatible Customers", "abstract": "A new variant of the classic capacitated facility location problem, which considers incompatibilities between customers, has recently been introduced in the literature. This problem captures the situation where given pairs of customers cannot be served by the same facility. Such a feature is crucial for many practical cases of location problems, such as the presence of hazardous or polluting materials and contention between competing costumers. In this paper, we propose a Large Neighborhood Search (LNS) method to solve this problem. Within the framework of LNS, we introduce three different destroy operators, which are combined in a hybrid manner, and we use an exact solver in the repair phase. Different algorithmic components are investigated for the design of LNS. The experimental analysis shows that our new method outperforms existing state-of-the-art metaheuristics, providing new best solutions for all available benchmark instances.", "date": "2026-05-27", "categories": []} {"id": "2605.28730", "title": "AlphaTransit: Learning to Design City-scale Transit Routes", "abstract": "Designing a transit network requires many sequential route extension decisions, but their quality is often visible only after the full network is assembled. This delayed-feedback challenge lies at the heart of the Transit Route Network Design Problem (TRNDP), where route interactions can be deceptive: an extension that appears useful locally can create transfer bottlenecks, produce redundant overlap, or reduce overall throughput. To guide route construction under delayed simulator feedback, we introduce AlphaTransit, a search-based planning framework for cityscale bus network design. AlphaTransit couples Monte Carlo Tree Search (MCTS) with a neural policy-value network: the policy proposes route extensions, the value estimates downstream design quality, and search uses these predictions to refine each decision. This provides decision-time lookahead during route construction without running simulator rollouts inside the search tree. We evaluate AlphaTransit on a new Bloomington TRNDP benchmark with realistic road topology and censusderived demand, under mixed and full transit demand settings. In the Bloomington network, AlphaTransit attains the highest service rate in both demand settings, reaching 54.6% and 82.1%, respectively. Relative to reinforcement learning without search, these correspond to 9.9% and 11.4% service rate gains; relative to MCTS without learned guidance, they correspond to 2.5% and 11.2% gains. These results suggest that coupling learned guidance with MCTS is more effective than using either approach alone for transit network design. Our code and data are publicly available in https://github.com/poudel-bibek/AlphaTransit.", "date": "2026-05-27", "categories": []} {"id": "2511.02525", "title": "An End-to-End Learning Approach for Solving Capacitated Location-Routing Problems", "abstract": "The capacitated location-routing problems (CLRPs) are classical problems in combinatorial optimization, which require simultaneously making location and routing decisions. In CLRPs, the complex constraints and the intricate relationships between various decisions make the problem challenging to solve. With the emergence of deep reinforcement learning (DRL), it has been extensively applied to address the vehicle routing problem and its variants, while the research related to CLRPs still needs to be explored. In this paper, we propose the DRL with heterogeneous query (DRLHQ) to solve CLRP and open CLRP (OCLRP), respectively. We are the first to propose an end-to-end learning approach for CLRPs, following the encoder-decoder structure. In particular, we reformulate the CLRPs as a markov decision process tailored to various decisions, a general modeling framework that can be adapted to other DRL-based methods. To better handle the interdependency across location and routing decisions, we also introduce a novel heterogeneous querying attention mechanism designed to adapt dynamically to various decision-making stages. Experimental results on both synthetic and benchmark datasets demonstrate superior solution quality and better generalization performance of our proposed approach over representative traditional and DRL-based baselines in solving both CLRP and OCLRP.", "date": "2025-11-04", "categories": []} {"id": "2604.04940", "title": "ReVEL: Multi-Turn Reflective LLM-Guided Heuristic Evolution via Structured Performance Feedback", "abstract": "Designing effective heuristics for NP-hard combinatorial optimization problems remains challenging and often requires substantial domain expertise. Recent LLM-guided evolutionary methods have shown promise for automated heuristic generation, but most existing approaches refine heuristics independently or through limited pairwise feedback. We propose ReVEL: Multi-Turn Reflective LLM-Guided Heuristic Evolution via Structured Performance Feedback, a framework for group-wise multi-turn heuristic refinement. ReVEL organizes heuristics into behavior-aware reflective groups, including similarity-driven groups for localized refinement and diversity-driven groups for exploratory search. Within each group, the LLM performs iterative multi-turn refinement using accumulated performance feedback, enabling related heuristics to be jointly analyzed and progressively improved across evolutionary iterations. Experiments on standard combinatorial optimization benchmarks show that ReVEL generally improves optimization performance over existing LLM-guided evolutionary baselines across multiple settings and LLM backbones. Additional analyses suggest that behavior-aware grouping contributes to more consistent refinement trajectories during iterative heuristic evolution.", "date": "2026-03-05", "categories": []} {"id": "2605.26763", "title": "Adversarial Training for Robust Coverage Network under Worst-case Facility Losses", "abstract": "The Maximal Covering Location-Interdiction Problem (MCLIP) is a classic bi-level optimization problem, which is fundamental to resilient infrastructure planning yet remains computationally intractable. Specifically, the upper level determines facility locations to maximize coverage, while the lower level executes worst-case interdiction to minimize the coverage. The strong coupling between the upper and lower levels, combined with their respective high combinatorial complexity, renders traditional methods ineffective. To bridge this gap, we propose a Dual-Agent Deep Reinforcement Learning (DADRL) framework based on adversarial learning, comprising a location agent corresponding to the upper level and an interdiction agent corresponding to the lower level. Our contributions are threefold: (1) The location agent is trained simultaneously against an evolving interdiction agent, making it effectively capture the dynamic competitive interplay between the upper and lower levels; (2) To fully exploit the learned capabilities of the interdiction agent, we propose a Surrogate-based Ensemble Inference Strategy that utilizes the trained interdiction agent as a high-fidelity surrogate to guide the decisions of location agent; (3) Extensive experiments on synthetic and real-world datasets demonstrate that our approach achieves superior computational efficiency while maintaining highly competitive solution quality compared to other baselines. Furthermore, our DADRL framework is model-agnostic to network structures, while its underlying adversarial learning paradigm demonstrates strong potential for solving other bi-level optimization problems.", "date": "2026-05-26", "categories": []} {"id": "2605.26776", "title": "Towards Generalization-Oriented Models for Vehicle Routing Problems with Mixture-of-Experts", "abstract": "In recent years, Deep Reinforcement Learning (DRL) has achieved substantial progress on Vehicle Routing Problems (VRPs). However, existing DRL-based methods are typically trained on instances generated from a uniform distribution, which limits their performance under real-world distribution shifts. In this paper, we aim to develop a generalization-oriented model that partitions the policy network into multiple modules and adaptively recombines modules to form specific policies during inference. Specifically, we propose Residual Refined Experts with Instance-level Gating (R2E-IG) to improve cross-distribution generalization. Our contributions are threefold: (1) We introduce a Residual Refined Expert (R2E) architecture that enhance expert expressiveness via residual refinement; (2) We design an instance-level gating mechanism that learns distribution-aware instance representations and routes inputs to suitable modules; (3) We propose a mixed-distribution training mechanism equipped with Dynamic Weight Adaption (DWA), which dynamically reweights training data from different distributions to emphasize more informative ones. Extensive experiments show that R2E-IG achieves competitive performance against state-of-the-art baselines on both in-distribution and out-of-distribution instances across synthetic and benchmark datasets. Moreover, R2E-IG is generic and can be easily integrated into existing DRL-based methods to further improve performance.", "date": "2026-05-26", "categories": []} {"id": "2405.01906", "title": "Instance-Conditioned Adaptation for Large-scale Generalization of Neural Routing Solver", "abstract": "The neural combinatorial optimization (NCO) method has shown great potential for solving routing problems of intelligent transportation systems without requiring expert knowledge. However, existing constructive NCO methods still struggle to solve large-scale instances, which significantly limits their application prospects. To address these crucial shortcomings, this work proposes a novel Instance-Conditioned Adaptation Model (ICAM) for better large-scale generalization of neural routing solvers. In particular, we design a simple yet efficient instance-conditioned adaptation function to significantly improve the generalization performance of existing NCO models with a small time and memory overhead. In addition, with a systematic investigation on the performance of information incorporation between different attention mechanisms, we further propose a powerful yet low-complexity instance-conditioned adaptation module to generate better solutions for instances across different scales. Extensive experimental results on both synthetic and benchmark instances show that our proposed method is capable of obtaining promising results with a very fast inference time in solving large-scale Traveling Salesman Problems (TSPs), Capacitated Vehicle Routing Problems (CVRPs), and Asymmetric Traveling Salesman Problems (ATSPs). Our code is available at https://github.com/CIAM-Group/ICAM.", "date": "2024-05-03", "categories": []} {"id": "2509.23413", "title": "URS: A Unified Neural Routing Solver for Cross-Problem Zero-Shot Generalization", "abstract": "Multi-task neural routing solvers have emerged as a promising paradigm for their ability to solve multiple vehicle routing problems (VRPs) using a single model. However, existing neural solvers typically rely on predefined problem constraints or require per-problem fine-tuning, which substantially limits their zero-shot generalization ability to unseen VRP variants. To address this critical bottleneck, we propose URS, a unified neural routing solver that achieves zero-shot generalization across a wide range of unseen VRPs with a single model. We propose a unified data representation (UDR) that replaces problem enumeration with data unification, thereby broadening the problem coverage and reducing reliance on domain expertise. In addition, we introduce a Mixed Bias Module (MBM) during encoding to improve node embeddings, which efficiently captures multiple priors inherent to various problems. On top of the UDR, we develop a problem-conditioned parameter generator to further improve zero-shot generalization. Extensive experiments show that URS consistently produces high-quality solutions for 110 VRP variants (including 99 unseen variants) while demonstrating impressive scalability to large-scale instances with up to 7000 nodes. To the best of our knowledge, URS is the first neural solver to handle over 100 VRP variants with a single model. Our code is available at https://github.com/CIAM-Group/URS.", "date": "2025-09-27", "categories": []} {"id": "2601.20539", "title": "PathWise: Planning through World Model for Automated Heuristic Design via Self-Evolving LLMs", "abstract": "Large Language Models (LLMs) have enabled automated heuristic design (AHD) for combinatorial optimization problems (COPs), but existing frameworks' reliance on fixed evolutionary rules and static prompt templates often leads to myopic heuristic generation, redundant evaluations, and limited reasoning about how new heuristics should be derived. We propose a novel multi-agent reasoning framework, referred to as Planning through World Model for Automated Heuristic Design via Self-Evolving LLMs (PathWise), which formulates heuristic generation as a sequential decision process over an entailment graph serving as a compact, stateful memory of the search trajectory. This approach allows the system to carry forward past decisions and reuse or avoid derivation information across generations. A policy agent plans evolutionary actions, a world model agent generates heuristic rollouts conditioned on those actions, and critic agents provide routed reflections summarizing lessons from prior steps, shifting LLM-based AHD from trial-and-error evolution toward state-aware planning through reasoning. Experiments across diverse COPs show that PathWise converges faster to better heuristics, generalizes across different LLM backbones, and scales to larger problem sizes.", "date": "2026-01-28", "categories": []} {"id": "2605.14559", "title": "PyCSP3-Scheduling: A Scheduling Extension for PyCSP3", "abstract": "PyCSP$^3$ provides a productive way to build constraint models for solving combinatorial constrained problems and export them to XCSP$^3$, preserving a complete separation between modeling and solving. However, it lacks native support for scheduling abstractions such as interval variables, sequence variables, and resource functions. As a result, scheduling models must be encoded with low-level integer variables and manual channeling constraints, even though PyCSP$^3$ already provides global constraints like NoOverlap and Cumulative on integer arrays. We present PyCSP$^3$ Scheduling, a library that adds scheduling abstractions to PyCSP$^3$ through 53 dedicated constraints and 27 expressions, and compiles them down to standard PyCSP$^3$/XCSP$^3$ constraints, maintaining the modeling/solving separation that underpins the PyCSP$^3$ ecosystem. On 261 paired instances across 17 model families (5 runs each), both formulations produce identical objectives on all 72 doubly-proved optimal pairs and nearly half of the families (8/17) remain structurally unchanged after compilation; however, runtime performance diverges across families, with clear gains on some (up to 5.8x) and regressions on others due to the overhead of compilation decompositions. Code and benchmarks are available at: https://github.com/sohaibafifi/pycsp3-scheduling", "date": "2026-05-14", "categories": []} {"id": "2605.23395", "title": "Convex Compositional Reasoning Models", "abstract": "Compositional energy-based models can generalize to larger combinatorial reasoning problems by reusing a learned factor energy across many local constraints. In our paper, we show that a key bottleneck in compositional reasoning is not composition itself, but the non-convex geometry of the learned energy landscape. To solve this problem, we introduce Convex Compositional Energy Minimization (CCEM), a framework that parameterizes each factor with an input-convex neural network and optimizes the composed energy over a tight convex relaxation of the feasible set. Because convexity is preserved under summation, the global relaxed objective remains convex, enabling deterministic projected first-order optimization. CCEM is trained in two stages: factor-level contrastive learning to shape local energy basins, followed by end-to-end refinement through an unrolled projected solver. Our experiments show that our models trained on small subproblems or a single problem size transfer to larger instances without retraining.", "date": "2026-05-22", "categories": []} {"id": "2605.23957", "title": "Low-Cost Labels, Reliable Choices: Rollout-Calibrated Hyper-Heuristics for Job Shop Scheduling", "abstract": "Learning-assisted hyper-heuristics can select among dispatching rules while preserving the feasibility and interpretability of constructive Job Shop Scheduling Problem (JSSP) heuristics. Their main computational cost lies in label generation rather than model fitting, since each supervised label usually requires rolling out candidate rules from a partial schedule. We study this label-cost problem together with a reliability problem: a learned selector should not switch away from a strong default rule unless the predicted gain is credible. The proposed selector uses regret-normalized rollout labels, a contextual KNN uncertainty estimate, and a gate that acts only when the predicted improvement exceeds an uncertainty-adjusted margin. We also vary rollout depth and breadth to measure the cost-quality trade-off. On synthetic JSSP instances, the gated selector achieves the lowest mean RPD among learned selectors, remains close to the best fixed dispatching rule, and reduces Random-HH mean RPD by more than an order of magnitude.", "date": "2026-05-11", "categories": []} {"id": "2605.24484", "title": "SPACE: Unifying Symmetric and Asymmetric Routing Problems for Generalist Neural Solver", "abstract": "Generalist neural routing solvers have shown great potential in solving diverse vehicle routing problems (VRPs) with a unified model. However, existing solvers are typically limited to symmetric settings or degrade in performance when switching to asymmetric settings due to input inconsistencies or inherent structural differences, substantially limiting their practicality in real-world scenarios that encompass both scenarios. To address this limitation, we define the spatial position of each node based on the relative distances to a specific set of pivots and further propose a Spatial Pivot-Aligned Coordinate-free Embedding (SPACE) framework that unifies node representation and solution generation across symmetric and asymmetric VRPs. Specifically, we construct a bidirectional Frechet representation using a novel furthest pivot sampling strategy to enable invariant node representations across distinct problem settings. Furthermore, we introduce a weight-decomposed adaptive decoding mechanism that decouples geometric perception from problem representations, mitigating the overfitting of constraint decisions to a specific geometry setting. Extensive experiments on 110 VRP variants, comprising 55 symmetric problems and their asymmetric counterparts, demonstrate that SPACE achieves promising zero-shot generalization in both symmetric and asymmetric VRPs.", "date": "2026-05-23", "categories": []} {"id": "2605.24502", "title": "Implicit Binarization via Complex Phase Dynamics in Combinatorial Optimization", "abstract": "We introduce a physics-inspired continuous relaxation framework that yields substantially improved solutions for NP-hard combinatorial optimization problems, including Quadratic Unconstrained Binary Optimization (QUBO), binary sparse coding, and planted-solution Ising models. By parameterizing discrete binary variables as continuous wave-like states on the complex unit circle, we inherently smooth highly non-convex energy landscapes. We show that representing binary variables as complex phases reveals an implicit regularization mechanism that promotes convergence toward discrete states. Extracting this mechanism yields significant improvements even within standard real-valued optimization frameworks, using this regularizer explicitly. Empirically, this regularization yields vastly higher ground-state convergence rates than standard real-valued alternatives. Our models achieved zero error in large-scale 160x160 QUBO tasks under severe noise (sigma=0.25), and outperformed traditional algorithms (OMP and LASSO) in underdefined sparse coding with perfect recovery at sigma=0.15. The solver's robustness was further validated by recovering exact ground-state configurations in 8 out of 11 rigorously engineered planted-solution benchmarks.", "date": "2026-05-23", "categories": []} {"id": "2605.24779", "title": "Complement Submodular Information Measures for Balanced and Robust Data Selection", "abstract": "Submodular optimization has become a fundamental paradigm for data selection, retrieval, summarization, and representation learning due to its ability to model coverage, diversity, and representativeness. However, classical submodular objectives optimize only the selected subset and do not explicitly preserve structural information between the selected subset and the remaining data. In many modern machine learning applications, including train/validation/test splitting, benchmark construction, and robust subset selection, the quality of a selection depends critically on preserving balanced structure across both the selected subset and its complement.\n In this work, we introduce Complement Submodular Information (CSI), a new class of complement-aware submodular objectives that quantify shared structural information between a subset and its complement. Our framework induces complement-aware variants of several classical submodular functions including Facility Location, Graph Cut, LogDet, Saturated Coverage, Set Cover, Probabilistic Set Cover, and Feature Based Functions. We analyze the theoretical properties of CSI objectives and show that they exhibit approximate monotonicity under bounded curvature conditions, leading to near-$(1-1/e)$ greedy approximation guarantees.\n Empirically, CSI objectives consistently outperform standard submodular objectives on robust hidden-slice-aware subset selection. In particular, CSI objectives significantly improve preservation of coherent rare/tail semantic structure while simultaneously suppressing noisy and isolated outliers, leading to substantially improved downstream predictive performance. Synthetic experiments further illustrate how different CSI instantiations capture complementary notions of representativeness, diversity, connectivity, and balanced neighborhood preservation.", "date": "2026-05-23", "categories": []} {"id": "2605.25235", "title": "Constraint-Anchored Attribution: Feasibility-Certified Counterfactuals and Bonferroni-PAC Sufficient Subsets for Neural CO Policies", "abstract": "We give an attribution method for neural combinatorial-optimisation (CO) policies that (i) decomposes a decision by constraint families via LP-relaxation duals, (ii) certifies counterfactuals through a combinatorial feasibility model (implemented as a CSP feasibility-decision model), and (iii) bounds the size of a PAC-sufficient explanation with a Bonferroni-corrected Hoeffding sufficient-subset test along a greedy ordering. Across three CO problems and three seeds, our LP-anchored $\\Lambda$-attribution matches the CF-derived signal at 96.5% on CVRPTW (n_cert=344) and 77.2% on the Orienteering Problem (n_cert=281) vs 75.0% and 35.2% for proxy gradient (paired diffs +0.215 and +0.420; McNemar exact $p \\le 10^{-14}$). In the rank-aligned regime of the Flexible Job-Shop Scheduling Problem, both backends agree on every CSP-certified flip (n_cert=59), confirming the no-gain prediction. Bonferroni-PAC subsets average 5.0 nodes per step ($M=70$, $\\varepsilon=\\delta=0.2$, $k_{\\max}=25$). Reference implementation: https://github.com/sohaibafifi/neuro-co-cax", "date": "2026-05-24", "categories": []} {"id": "2605.25551", "title": "Learning Permutation from Structure Without Supervision", "abstract": "Many learning problems require uncovering a hidden ordering that reveals structure in unordered data, such as monotonicity in sorting or spatial continuity in jigsaw reconstruction. In these settings, permutations can be learned as latent operators by optimizing objectives defined directly on the reordered output, often without access to ground-truth orderings. Differentiable relaxations such as Gumbel-Sinkhorn make this approach practical by approximating permutation matrices with doubly stochastic matrices. However, learning from structure without supervision induces a non-uniform uncertainty: some assignments become confident early, while others remain ambiguous. Existing methods control this process using a single global temperature, forcing all assignments to sharpen or diffuse simultaneously and leading to instability at scale. We introduce an entropy-adaptive formulation of Gumbel-Sinkhorn that locally modulates temperature based on assignment uncertainty. This allows confident assignments to discretize early while preserving exploration where uncertainty remains. Across sorting and jigsaw reconstruction tasks and in routing-style settings, adaptive entropy control improves training stability and final permutation quality relative to fixed-temperature baselines, particularly as problem size and assignment ambiguity increase.", "date": "2026-05-25", "categories": []} {"id": "2605.22876", "title": "WeCon: An Efficient Weight-Conditioned Neural Solver for Multi-Objective Combinatorial Optimization Problems", "abstract": "Existing neural solvers for Multi-Objective Combinatorial Optimization Problems (MOCOPs) commonly adopt decomposition-based strategies that scalarize an MOCOP into multiple subproblems associated with distinct weight vectors. However, they either inject weights only once during decoding, limiting weight-conditioned context modeling, or primarily during encoding, causing weight-signal dilution during decoding. Moreover, preference optimization methods rely on purely random sampling to construct solution pairs for training solvers, which often produces less informative pairs and thus leads to low training effectiveness. To better address these limitations, we propose an efficient Weight-Conditioned neural solver (WeCon). Specifically, we design an encoder layer with three attention blocks and our proposed Gated Residual Fusion (GRF) block to facilitate harmonious interaction between instance features and weights, thereby generating informative weight-conditioned context. We further introduce a plug-and-play Residual Fusion (RF) block in the decoder to alleviate weight-signal dilution. Finally, we propose Efficient Preference Optimization (EPO), which constructs high-quality solutions, thereby generating more informative pairs to improve training effectiveness. Experiments on four MOCOP variants across different problem scales and distribution patterns demonstrate that WeCon achieves HyperVolume (HV) values comparable to SOTA solver POCCO-W, while reducing inference time by 40%. Ablation studies validate the contributions of all designs.", "date": "2026-05-20", "categories": []} {"id": "2605.23569", "title": "CP or DP? Why Not Both: A Case Study in the Partial Shop Scheduling Problem", "abstract": "Dynamic Programming (DP) and Constraint Programming (CP) are well-established paradigms for solving combinatorial optimization problems. Usually, these two approaches are used separately. This paper aims to show that the two can be combined effectively and elegantly, with DP serving as the primary search framework and CP used as a subroutine to leverage global constraint propagation. This paper presents such an approach for the Partial Shop Scheduling Problem (PSSP), for which a pure DP method has previously been proposed, and efficient CP filtering algorithms are available. The PSSP is a general scheduling problem where each job consists of a set of operations with arbitrary precedence constraints. The approach is flexible enough to accommodate anytime DP strategies, such as anytime column search, whereas the original DP algorithm operated in a strictly layer-wise manner. Moreover, the flexibility of the CP modeling makes it straightforward to incorporate arbitrary precedence constraints. As a result, the model naturally handles any precedence graph and even enables the design of a Large Neighborhood Search (LNS) scheme, in which the DP model is reused, and partial-order schedules are imposed across restarts to improve the incumbent solution. While not competitive with state-of-the-art pure CP solvers for this specific problem, our primary contribution is demonstrating the viability of this hybrid integration.", "date": "2026-05-22", "categories": []} {"id": "2605.23592", "title": "Solving the Aircraft Disassembly Scheduling Problem", "abstract": "Dismantling aircrafts reaching their end of life is a complex endeavour that is necessary in terms of sustainability but yields small income margins for air transport companies. An efficient scheduling of the disassembly procedure is thus crucial to ensure the profitability of the process and incentivize practice. This is a large scheduling problem that involves thousands of tasks and many different constraints: Extracting parts that are destined to be reused requires technicians with specific certifications and equipment. Extraction operations might be subject to precedence relations. Furthermore, the aircraft must be kept balanced during the whole process. Finally, some of the locations of the aircraft have a limited space that caps the number of technicians able to work there concurrently. This article presents the problem in details and proposes two approaches to solve the problem: a Constraint Programming model and a MIP model. The models are tested on instances of varying sizes involving up to 1450 tasks, which are based on real operational data provided by an industrial partner.", "date": "2026-05-22", "categories": []} {"id": "2605.21519", "title": "Neural Acceleration for Graph Partitioning", "abstract": "Graph Partitioning is a critical problem in numerous scientific and engineering domains including social network analysis, VLSI design, and many more. Spectral methods are known to produce quality partitions while minimizing edge cuts for a wide range of problems. However, the computational cost associated with the calculation of the Fiedler vector, an eigenvector associated with the second smallest eigenvalue of the graph Laplacian, remains a significant bottleneck due to memory issues and computational costs. In this paper, we present an accelerated approach to spectral bisection partitioning by replacing the traditional eigenvalue calculation with a simple artificial neural network model to approximate the Fiedler vector. We demonstrate that our approach achieves partitioning quality comparable to spectral bisection while significantly reducing the computational overhead, making it more scalable and efficient for large-scale problems", "date": "2026-05-18", "categories": []} {"id": "2605.22188", "title": "From Sequential Nodes to GPU Batches: Parallel Branch and Bound for Optimal $k$-Sparse GLMs", "abstract": "GPUs have significantly accelerated first-order methods for large-scale optimization, especially in continuous optimization. However, this success has not transferred cleanly to problems with discrete variables, combinatorial structure, and nonlinear objectives, such as certifying optimal solutions for cardinality-constrained generalized linear models. Major challenges include the sequential processing of heterogeneous nodes in branch and bound (BnB) and frequent data movement between the CPU and GPU. We propose a simple, generic, and modular CPU--GPU framework that processes multiple BnB nodes in batches on GPUs. The framework is built around a small set of GPU-efficient routines and uses padding together with lightweight custom kernels to handle irregular node data structures. Experiments show one to two orders of magnitude speedups and zero optimality gap on challenging instances. The framework can also be extended to collect the entire Rashomon set, enabling downstream statistical analysis such as variable-importance analysis and model selection under secondary user-specific measures (e.g., AUC in classification).", "date": "2026-05-21", "categories": []} {"id": "2605.22773", "title": "Deep Reinforcement Learning for Flexible Job Shop Scheduling with Random Job Arrivals", "abstract": "The Flexible Job Shop Scheduling Problem (FJSP) is the optimal allocation of a set of jobs to machines. Two primary challenges persist in FJSP: the unpredictable arrival of future jobs and the combinatorial complexity of the problem, rendering it intractable for conventional mixed-integer linear programming solvers. This paper proposes an event-based \\gls{DRL} approach to solve FJSP with random job arrivals. Specifically, we employ the Proximal Policy Optimization algorithm and use lightweight Multi-Layer Perceptrons to train the \\gls{DRL} agent for minimizing the total completion time of all jobs. We design the state representation to be directly accessible from the environment, and limit the learning agent to selecting from among a set of well-established dispatching rules. Simulations show that our \\gls{DRL} approach outperforms any of the individual dispatching rules on datasets with varying heterogeneity and job arrival rates. We benchmark our \\gls{DRL} against an arrival-triggered mixed-integer linear programming solution and show that our method achieves good performance especially when the datasets are heterogeneous.", "date": "2026-05-21", "categories": []} {"id": "2605.18623", "title": "An Approximation Algorithm for Graph Label Selection", "abstract": "In the graph label selection problem, one is given an $n$-vertex graph and a budget $k$, and seeks to select $k$ vertices whose labels enable accurate prediction of the labels on the remaining vertices. This problem formalizes distilling a small representative set from the whole graph. We present the first $\\tilde{O}(\\log^{1.5} n)$-approximation algorithm for graph label selection under the standard budget constraint. Prior work either relies on resource augmentation, allowing substantially more than $k$ labeled vertices, or consists primarily of heuristics without provable guarantees. Finally, we demonstrate that practical heuristic variants of our algorithm scale to significantly larger graphs than previous methods, while essentially retaining their quality.", "date": "2026-05-18", "categories": []} {"id": "2605.20328", "title": "Targeting Clause Type Distributions: a Picklock for Random Satisfiability Problems", "abstract": "Optimization problems such as the NP-complete 3-SAT provide an important benchmark for the difficult task of finding ground-states in strongly correlated many-body systems with rugged energy landscapes. The study of random 3-SAT problems as Ising spin Hamiltonians in statistical physics has yielded major insights including the existence of a satisfiability phase transition, and the prediction of a critical parameter line of particularly hard instances. Yet, progress on solving those instances has been scarce for several decades. Here, introducing the Target-SAT (TSAT) algorithm, we roughly triple the tractable problem sizes in the hardest regime, with an even greater improvement in a vast range of neighboring regions. By leveraging statistical information hidden in the combinatorial constraints of the problem, TSAT is actively guided in its stochastic local search toward a target within the relevant parameter space. Our analysis also explains why established local search algorithms are limited to relatively small system sizes due to a vast low-energy trap. Furthermore, we characterize the aforementioned critical line in terms of a dominant additional complexity barrier, whose exponential scaling is quickly overcome by TSAT only in the surrounding parameter space. With TSAT, the lead in solving the hardest known random satisfiability problems returns to the realm of stochastic local search algorithms.", "date": "2026-05-19", "categories": []} {"id": "2605.20618", "title": "COAgents: Multi-Agent Framework to Learn and Navigate Routing Problems Search Space", "abstract": "Although Vehicle Routing Problems (VRP) are essential to many real-world systems, they remain computationally intractable at scale due to their combinatorial complexity. Traditional heuristics rely on handcrafted rules for local improvements and occasional \\textit{jumps} to escape local minima, but often struggle to generalize across diverse instances. We introduce \\textbf{COAgents}, a cooperative multi-agent framework that models the search process as a graph: nodes represent solutions, and edges correspond to either local refinements or large perturbations for diversification (i.e., jumps). A \\textit{Partial Search Graph} (PSG) is dynamically constructed during search, enabling COAgents to train a Node Selection Agent and a Move Selection Agent to guide intensification, and a Jump Agent to trigger well-timed explorations of new regions. Unlike end-to-end learning approaches, COAgents cleanly separates problem-agnostic search control from compact domain-specific encoding, facilitating adaptability across tasks. Extensive experiments on the CVRP and VRPTW benchmarks show that COAgents remains competitive with several learn-to-search baselines on CVRP and sets a new state of the art among learning-based methods on the more challenging VRPTW instances, reducing the gap to the best-known solutions by 14\\% at $N\\!=\\!100$ and 44\\% at $N\\!=\\!50$ relative to the strongest neural solver (POMO), and by 21\\% and 40\\% respectively relative to ALNS.\n Code is available at https://github.com/mahdims/COAgents.", "date": "2026-05-20", "categories": []} {"id": "2404.16676", "title": "Multilayer Correlation Clustering", "abstract": "We establish Multilayer Correlation Clustering, a novel generalization of Correlation Clustering to the multilayer setting. In this model, we are given a series of inputs of Correlation Clustering (called layers) over the common set $V$ of $n$ elements. The goal is to find a clustering of $V$ that minimizes the $\\ell_p$-norm ($p\\geq 1$) of the multilayer-disagreements vector, which is defined as the vector (with dimension equal to the number of layers), each element of which represents the disagreements of the clustering on the corresponding layer. For this generalization, we first design an $O(L\\log n)$-approximation algorithm, where $L$ is the number of layers. We then study an important special case of our problem, namely the problem with the so-called probability constraint. For this case, we first give an $(\\alpha+2)$-approximation algorithm, where $\\alpha$ is any possible approximation ratio for the single-layer counterpart. Furthermore, we design a $4$-approximation algorithm, which improves the above approximation ratio of $\\alpha+2=4.5$ for the general probability-constraint case. Computational experiments using real-world datasets support our theoretical findings and demonstrate the practical effectiveness of our proposed algorithms.", "date": "2024-04-25", "categories": []} {"id": "2407.17200", "title": "Generalization Bounds of Surrogate Policies for Combinatorial Optimization Problems", "abstract": "Many real-world decision problems require solving, again and again, combinatorial optimization instances drawn from a common distribution. A recent line of structured learning methods exploits this regularity by learning policies that pair a statistical model with a tractable combinatorial oracle, instead of solving each instance independently. Training such policies is notoriously difficult, however: the resulting empirical risk is piecewise constant in the model parameters, which hinders gradient-based optimization, and only a few theoretical guarantees have been provided so far. We address this issue by analyzing smoothed (perturbed) policies: adding controlled random perturbations to the direction used by the linear oracle yields a differentiable surrogate risk and improves generalization. Our main contribution is a generalization bound that decomposes the excess risk into $(\\mathit{i})$ perturbation bias, $(\\mathit{ii})$ statistical estimation error, and $(\\mathit{iii})$ optimization error. The perturbation bias is controlled by the \\emph{fan-crossing probability}, a new geometric quantity measuring the likelihood that a perturbation changes the oracle solution. We introduce two complementary conditions to bound it--the \\emph{Uniformly Bounded Density} (UBD) property, yielding a sharp ${O}(\\lambda)$ bias, and the weaker \\emph{Uniform Weak moment} (UW) property, yielding a sub-linear bound--both capturing the geometric interaction between the statistical model and the normal fan of the feasible polytope. The statistical estimation error is controlled via a uniform deviation bound over the policy class, with rate ${O}(1/(\\lambda\\sqrt{n}))$ that scales inversely in the smoothing parameter. Concerning the optimization error, we exploit kernel Sum-of-Squares methods to mitigate the curse of dimensionality of global optimization.", "date": "2024-07-24", "categories": []} {"id": "2510.12278", "title": "Quantum Annealing for Staff Scheduling in Educational Environments", "abstract": "We address a novel staff allocation problem that arises in the organization of collaborators among multiple school sites and educational levels. The problem emerges from a real case study in a public school in Calabria, Italy, where staff members must be distributed across kindergartens, primary, and secondary schools under constraints of availability, competencies, and fairness. To tackle this problem, we develop an optimization model and investigate a solution approach based on quantum annealing. Our computational experiments on real-world data show that quantum annealing is capable of producing balanced assignments in short runtimes. These results provide evidence of the practical applicability of quantum optimization methods in educational scheduling and, more broadly, in complex resource allocation tasks.", "date": "2025-10-14", "categories": []} {"id": "2604.03419", "title": "Adaptive Threshold-Driven Continuous Greedy Method for Scalable Submodular Optimization", "abstract": "Submodular maximization under matroid constraints is a fundamental problem in combinatorial optimization with applications in sensing, data summarization, active learning, and resource allocation. While the Sequential Greedy (SG) algorithm achieves only a $\\frac{1}{2}$-approximation due to irrevocable selections, Continuous Greedy (CG) attains the optimal $\\bigl(1-\\frac{1}{e}\\bigr)$-approximation via the multilinear relaxation, at the cost of a progressively dense decision vector that forces agents to exchange feature embeddings for nearly every ground-set element. We propose \\textit{ATCG} (\\underline{A}daptive \\underline{T}hresholded \\underline{C}ontinuous \\underline{G}reedy), which gates gradient evaluations behind a per-partition progress ratio $\\eta_i$, expanding each agent's active set only when current candidates fail to capture sufficient marginal gain, thereby directly bounding which feature embeddings are ever transmitted. Theoretical analysis establishes a curvature-aware approximation guarantee with effective factor $\\tau_{\\mathrm{eff}}=\\max\\{\\tau,1-c\\}$, interpolating between the threshold-based guarantee and the low-curvature regime where \\textit{ATCG} recovers the performance of CG. This shows that the problem structure, as captured by curvature, determines the amount of coordination and communication required to approach full-CG performance. Experiments on a class-balanced prototype selection problem over a subset of the CIFAR-10 animal dataset show that \\textit{ATCG} achieves objective values comparable to those of the full CG method while substantially reducing communication overhead through adaptive active-set expansion.", "date": "2026-04-03", "categories": []} {"id": "2605.17539", "title": "Memory-Guided Tree Search with Cross-Branch Knowledge Transfer for LLM Solver Synthesis", "abstract": "Combinatorial optimization (CO) underlies decision-making from logistics to chip design, where infeasible solutions are operationally unusable and small quality gains translate into substantial economic value. Recent work uses large language models (LLMs) to automate solver synthesis: generating executable solver programs from natural-language specifications. However, existing tree-search and evolutionary agents refine candidate trajectories in parallel without explicit knowledge transfer, reintroducing the same constraint violations and converging on similar algorithm families. We introduce MEMOIR, a memory-guided tree-search framework with a two-level memory hierarchy: branch-local memory preserves execution-grounded refinement details within a branch as it iterates on a single algorithmic design, while global memory stores compressed algorithmic and failure-mode summaries across branches. A reflection step at branch termination distills these summaries, enabling cross-branch transfer without polluting future contexts with low-level debugging traces. Across seven CO problems spanning scheduling, routing, packing, and geometric design, MEMOIR achieves 96.7% solution validity (a 9.2 point gap over the strongest baseline) and improves the average normalized score by 7.3 points at matched per-method execution budget. Over three independent runs on four problems, MEMOIR's run-to-run validity standard deviation is more than an order of magnitude below that of every baseline we evaluated in this setting, suggesting that memory-guided exploration yields consistent improvements rather than reflecting sampling variance.", "date": "2026-05-17", "categories": []} {"id": "2605.19119", "title": "GOAL: Graph-based Objective-Aligned Diffusion Solvers for Dynamic Multi-Objective Optimization", "abstract": "Existing neural combinatorial optimization solvers frame solution search as imitation of optimal decisions, inherently limiting their utility to single-objective minimization and static constraints. We propose GOAL, a conditioned diffusion solver over relational graph representations that enables controllable decision generations by conditioning on human-specified objectives. We introduce a heterogeneous graph encoding in which distinct edge types, corresponding to different classes of constraints, define the message passing structure of the graph neural network, which allows information to propagate selectively according to the ontology of each constraint. GOAL is instantiated and evaluated on three canonical scheduling benchmarks of various constraint complexity: the Flow Shop Problem (FSP), the Job Shop Scheduling Problem (JSP), and the Flexible Job Shop Scheduling Problem (FJSP). Generalization is demonstrated across structurally distinct constraint regimes and problem types without architectural modification. On all three benchmarks, GOAL achieves 100% solution feasibility and near-zero MAPE (below 0.20%) on multiple objectives for problem sizes up to 20 jobs and 60 operations, outperforming NSGA-II and MOEA/D in both solution quality and inference speed by up to 25x.", "date": "2026-05-18", "categories": []} {"id": "2605.19721", "title": "Projecting Latent RL Actions: Towards Generalizable and Scalable Graph Combinatorial Optimization", "abstract": "Graph combinatorial optimization (GCO) has attracted growing interest, as many NP-hard problems naturally admit graph formulations, yet their combinatorial explosion renders exact methods computationally intractable. Recent advances in Reinforcement Learning (RL) combined with Graph Neural Networks (GNNs) have significantly improved learning-based GCO solvers. However, existing approaches face limitations in both generalization across diverse graph instances and computational scalability as action spaces grow. To address both challenges, we introduce projection agents, a novel RL-GCO approach that operates directly in a continuous GNN-based action embedding space, predicting a desired latent action in a single forward pass and subsequently decoding it into a valid discrete action. Additionally, we enable fair comparison across RL methods through a shared embedding space for both observations and actions. Across diverse benchmarks, our approach achieves up to 16.2x faster inference and up to 40% better generalization than existing solutions using only simple nearest-neighbor decoding, while opening the door to strong RL performance in super-linear decision spaces with multiple interdependent variables. Finally, we release LaGCO-RL, a Python library that automates latent action-space construction and supports existing RL-GCO solutions, promoting reproducibility and adaptation to new GCO benchmarks.", "date": "2026-05-19", "categories": []} {"id": "2605.19975", "title": "Learning with Foresight: Enhancing Neural Routing Policy via Multi-Node Lookahead Prediction", "abstract": "Neural policies have shown promise in solving vehicle routing problems due to their reduced reliance on handcrafted heuristics. However, current training paradigms suffer from a fundamental limitation: they primarily focus on next-node prediction for solution construction, resulting in myopic decision-making that undermines long-horizon planning capacity. To this end, we introduce Multi-node Lookahead Prediction (MnLP), a novel training strategy that extends the supervised learning paradigm to predict multiple future nodes simultaneously. We incorporate causal and discardable MnLP modules that operate exclusively during training, facilitating models to anticipate multi-step decisions while preserving inference-time efficiency. By incorporating multi-depth auxiliary supervision into the loss function, MnLP equips neural policies with the ability of long-range contextual understanding. Experimentally, MnLP outperforms existing training methods, improving the generalization capability of neural policies across various problem sizes, distributions, and real-world benchmarks. Moreover, MnLP can be seamlessly integrated into diverse neural architectures without introducing additional inference overhead.", "date": "2026-05-19", "categories": []} {"id": "2605.20074", "title": "Towards Distillation Guarantees under Algorithmic Alignment for Combinatorial Optimization", "abstract": "Distillation transfers knowledge from a large model trained on broad data to a smaller, more efficient model suitable for deployment. In structured prediction settings, prior knowledge about the task can guide the choice of a target architecture that is algorithmically aligned with the underlying problem. Building on recent learning-theoretic analyses of decision-tree (DT) distillation (Boix-Adsera, 2024), we study when distillation succeeds for combinatorial optimization tasks. We focus on the case where the target model is a graph neural network whose architecture is aligned with a dynamic programming (DP) algorithm for the task. Assuming that the source model is sufficiently rich, formalized through the linear representation hypothesis (LRH) (Elhage et al., 2022; Park et al., 2024), we show that the distillation problem can be solved efficiently in the complexity parameters of the DP transition function, represented as a DT. Our results provide a rigorous sufficient condition for successful distillation in the flavour of algorithmic alignment.", "date": "2026-05-19", "categories": []} {"id": "2605.14624", "title": "An Amortized Efficiency Threshold for Comparing Neural and Heuristic Solvers in Combinatorial Optimization", "abstract": "A common critique of neural combinatorial-optimization solvers is that they are less energy-efficient than CPU metaheuristics, given the operational energy cost of training them on GPUs. This paper examines the inferential step from \"training is expensive\" to \"neural solvers are net-inefficient\", which is where the critique actually goes wrong. Training the network costs a large fixed amount of GPU energy; running the metaheuristic costs a small amount of CPU energy on every instance, repeated as long as the solver is deployed. The two are not commensurable until a deployment volume is fixed. We define the Amortized Efficiency Threshold (AET) as the deployment volume above which a neural solver breaks even with a heuristic baseline in total energy or carbon, under an explicit constraint on solution quality. We show that the cumulative-energy ratio between the two solvers tends to a constant strictly below one whenever the network wins per instance, and that this limit does not depend on how the training cost was measured. An embodied-carbon term amortizes hardware fabrication symmetrically on both sides. We instantiate the framework on the CVRP environment at n=50 customers with the attention-based autoregressive solver of Kool et al. (2019), trained for 100 epochs on 20,000 instances over five random seeds, and HGS via PyVRP as the heuristic baseline. The measured operational crossover sits near 4.56e3 deployed instances at the median of a six-point baseline-budget sweep; the per-instance neural-to-heuristic ratio is 2.29e-3. The contribution is the framework, the open instrumentation, and the end-to-end measurement protocol. Code and benchmark pipeline are available at https://github.com/sohaibafifi/aet.", "date": "2026-05-14", "categories": []} {"id": "2605.17137", "title": "Latent Heuristic Search: Continuous Optimization for Automated Algorithm Design", "abstract": "The integration of Large Language Models (LLMs) into evolutionary frameworks has established a new paradigm for automated heuristic discovery. Despite their promise, these methods typically search in the discrete space of program syntax, relying on stochastic sampling to navigate a highly non-convex optimization landscape. This work proposes a continuous heuristic discovery framework that shifts optimization to a learned latent manifold. We employ an encoder to map discrete programs into continuous embeddings and train a differentiable surrogate model to predict performance, enabling gradient-based search. To regularize the optimization trajectory, an invertible normalizing flow maps these embeddings to a structured Gaussian prior, where we perform gradient ascent. The resulting optimized latent vectors are projected through a learned mapper into soft prompts, which condition a frozen LLM to synthesize novel executable heuristics. We evaluate the proposed method on the Traveling Salesman Problem (TSP), the Capacitated Vehicle Routing Problem (CVRP), the Knapsack Problem (KSP), and Online Bin Packing (OBP). Empirical results demonstrate that continuous latent-space optimization achieves performance competitive with state-of-the-art discrete evolutionary baselines while offering a complementary methodological alternative for automated algorithm design. The implementation code is available at \\url{https://github.com/cheikh025/LHS}.", "date": "2026-05-16", "categories": []} {"id": "2605.18094", "title": "Learning to Solve Compositional Geometry Routing Problems", "abstract": "We study the Compositional Geometry Routing Problem (CGRP), a unified superclass of traditional routing problems that covers point-only, line-only, area-only, and arbitrary hybrid task geometries, providing a broad abstraction for real-world routing scenarios. Beyond standard point-based routing, CGRP with non-point tasks can be inherently asymmetric, tightly coupled travel routes with the intrinsic path, and enlarges the action space with numerous feasible yet often irrelevant options, thereby posing significant challenges for both representation learning and decision-making. To address these challenges, we propose DiCon, a differential attention-assisted solver with contrastive learning, as a plug-and-play framework that tackles the problem from two complementary angles. First, we introduce a differential attention mechanism that actively suppresses the probability mass on less competitive candidate actions. Second, we design a double-level contrastive learning objective to promote robust global instance representations and regularize geometry-aware task representations. Extensive experiments demonstrate that DiCon achieves strong performance, broad versatility, and superior generalization across diverse CGRP instances with different compositions.", "date": "2026-05-18", "categories": []} {"id": "2605.18374", "title": "Beyond Inference-Time Search: Reinforcement Learning Synthesizes Reusable Solvers", "abstract": "Large language models (LLMs) typically approach combinatorial optimization as an inference-time procedure, solving each instance separately through sampling, search, or repeated prompting. We ask whether reinforcement learning can instead shift part of this reasoning cost into the weights of a code LLM, so that the model synthesizes a reusable solver for an entire problem family. We study this question on Synergistic Dependency Selection (SDS), a controlled variant of constrained Quadratic Knapsack designed to expose a specific failure mode: local signals and strict feasibility constraints make greedy heuristics attractive but unreliable. Under identical scaffolding, Best-of-64 base-model sampling saturates at an approximately 28.7% gap to the global Virtual Best Solver (VBS); code audits show that the base model often retrieves Simulated Annealing templates but misimplements the Metropolis acceptance rule. We fine-tune Qwen2.5-Coder-14B-Instruct with Group Relative Policy Optimization (GRPO) using a feasibility-gated reward and light structural scaffolding. The resulting policy converges to a constraint-aware Simulated Annealing template in 99.8% of feasible SDS outputs, achieves a 5.0% gap to that VBS, and is 91 times cheaper in post-generation execution/search cost than cumulative Best-of-64 evaluation. A compile-once check shows that one best frozen solver per seed remains highly competitive when reused unchanged across the SDS test set, while an additional-domain evaluation on Job Shop Scheduling provides narrower but positive evidence that the scaffold transfers beyond SDS. Negative ablations reveal the limits of this recipe: standard stabilizers degrade performance, a soft feasibility gate fails, and results remain sensitive to reward normalization and domain-specific design choices.", "date": "2026-05-18", "categories": []} {"id": "2605.18454", "title": "Scheduling That Speaks: An Interpretable Programmatic Reinforcement Learning Framework", "abstract": "Deep reinforcement learning (DRL) has recently emerged as a promising approach to solve combinatorial optimization problems such as job shop scheduling. However, the policies learned by DRL are typically represented by deep neural networks (DNNs), whose opaque neural architectures and non-interpretable policy decisions can lead to critical trust and usability concerns for human decision makers. In addition, the computational requirements of DNNs can further hinder practical deployment in resource constrained environments. In this work, we propose ProRL, a novel interpretable programmatic reinforcement learning framework that achieves high-performance scheduling with human-readable and editable programmatic policies (i.e., programs). We first introduce a domain-specific language for scheduling (DSL-S) to represent scheduling strategies as structured programs. ProRL then explores the program space defined by DSL-S using local search to identify incomplete programs, which are subsequently completed by learning their parameters via Bayesian optimization. ProRL learns which scheduling heuristic rules to select, and hence, it naturally incorporates existing heuristics already used in industrial scenarios. Experiments on widely used benchmark instances demonstrate the strong performance of ProRL against existing heuristics and DRL baselines. Furthermore, ProRL performs well under strongly constrained computational resources, such as training with only 100 episodes. Our code is available at https://github.com/HcPlu/ProRL.", "date": "2026-05-18", "categories": []} {"id": "2508.17218", "title": "Generalized Policy Gradient with History-Aware Decision Transformer for Reliable Routing over Graph Signals", "abstract": "Reliable path planning in stochastic transportation networks requires decisions that account for uncertain and correlated travel times on irregular road graphs, rather than only minimizing expected delay. Such networks exhibit strong spatial-temporal coupling, where link travel times evolve as stochastic processes over graph edges, making the problem inherently sequential under uncertainty. Existing stochastic on-time arrival (SOTA) methods primarily depend on the current node and remaining budget, which limits their ability to exploit trajectory-level temporal structure and history-dependent correlations. This work proposes GPG-HT, a history-aware graph-signal policy framework that integrates a Decision Transformer with generalized policy gradient optimization for reliable routing. By attending to historical node-edge-time observations, GPG-HT captures non-Markovian spatial-temporal dependencies and enables context-aware decision making under uncertainty. Experiments on the Sioux Falls and Anaheim networks demonstrate consistent gains in on-time arrival probability over representative optimization and reinforcement learning baselines.", "date": "2025-08-24", "categories": []} {"id": "2605.15983", "title": "Petri Net Induced Heuristic Search for Resource Constrained Scheduling", "abstract": "We formulate the Resource-Constrained Project Scheduling Problem (RCPSP) as optimal search over the reachability graph of a Timed Transition Petri Net with Resources, using relative-delay tokens so that scheduling decisions correspond to transition firings in the induced state space. We solve the resulting problem with $A^*$ guided by a heuristic that combines Critical Path and resource-based lower bounds, and prove that it is consistent under our token-based time semantics. Experiments on the PSPLIB benchmarks show that the approach outperforms strong exact Mixed-Integer Linear Programming (MIP) baselines (SCIP, CBC) in both success rate and solve time. Per-instance analysis shows that heuristic search and MIP degrade along independent axes, resource tightness for $A^*$ and formulation size for MIP, with resource strength mediating which solver benefits from scale.", "date": "2026-05-15", "categories": []} {"id": "2605.13849", "title": "Mixed Integer Goal Programming for Personalized Meal Optimization with User-Defined Serving Granularity", "abstract": "Determining what to eat to satisfy nutritional requirements is one of the oldest optimization problems in operations research, yet existing formulations have two persistent limitations: continuous variables produce impractical fractional servings (1.7 eggs, 0.37 bananas), and hard nutrient constraints cause infeasibility when targets conflict. A systematic review of 56 diet optimization papers found that none combine integer programming with goal programming to address both issues. We propose Mixed Integer Goal Programming (MIGP) for personalized meal optimization. The formulation uses integer variables for practical serving counts and goal programming deviations for soft nutrient targets, with inverse-target normalization to balance multi-nutrient optimization. Per-food serving granularity allows natural units (one egg, one tablespoon of oil) without post-hoc rounding. We characterize the integrality gap in the goal programming context and identify a deviation absorption property: GP deviation variables buffer the cost of requiring integer servings, making the gap structurally smaller than in hard-constraint MIP. For meals with 15+ foods, the integer solution matches the continuous optimum in every benchmark instance. A computational evaluation across 810 instances (30 USDA foods, 9 configurations, 3 methods) shows MIGP finds strictly better solutions than GP with post-hoc rounding in 66% of cases (never worse) while maintaining 100% feasibility; hard-constraint IP achieves only 48%. Solve times stay under 100 ms for typical meal sizes using the open-source HiGHS solver. The implementation is available as an open-source Python module integrated into an interactive meal planning application.", "date": "2026-03-12", "categories": []} {"id": "2605.14141", "title": "Distribution-Aware Algorithm Design with LLM Agents", "abstract": "We study learning when the learned object is executable solver code rather than a predictor. In this setting, correctness is not enough: two solvers may both return valid solutions on the deployment distribution while differing substantially in runtime. Given samples from an unknown task distribution, the learner returns code evaluated on fresh instances by both solution quality and execution time. Our central abstraction is a \\emph{solver hint}: reusable structure inferred from samples and compiled into specialized solver code. We prove that the empirically fastest sample-consistent solver from a fixed library generalizes in both correctness and runtime, and that statistically identifiable hints can be recovered and compiled from polynomially many samples.\n Empirically, we instantiate the framework with LLM code agents on \\(21\\) structured combinatorial-optimization target distributions across seven problem classes. The synthesized solvers reach mean normalized quality \\(0.971\\), improve by \\(+0.224\\) over the average heuristic pool and by \\(+0.098\\) over the highest-quality heuristic, and are \\(336.9\\times\\), \\(342.8\\times\\), and \\(16.1\\times\\) faster than the quality-best heuristic, Gurobi, and the selected time-limited exact backend, respectively. On released PACE 2025 Dominating Set private instances, the synthesized solver is valid on all \\(100\\) graphs and runs about two orders of magnitude faster than top competition solvers, with a moderate quality gap. Inspection shows that many gains come from changing the computational scale: replacing ambient exponential search or general-purpose optimization with compiled distribution-specific computation.", "date": "2026-05-13", "categories": []} {"id": "2605.14195", "title": "Stochastic Matching via Local Sparsification", "abstract": "The classic online stochastic matching problem typically requires immediate and irrevocable matching decisions. However, in many modern decentralized systems such as real-time ride-hailing and distributed cloud computing, the primary bottleneck is often local communication bandwidth rather than the timing of the match itself. We formalize this challenge by introducing a two-stage local sparsification framework. In this setting, arriving requests must prune their realized compatibility sets to a strict budget of $k$ edges before a central coordinator optimizes the global matching. This creates a \"middle ground\" between local information constraints and global optimization utility.\n We propose a local selection strategy, parametrized by a fractional solution of the expected instance. Theoretically, we quantify the approximation ratio as a function of the solution's {\\em spread}. We prove that under sufficient spread, our sparsifier globally preserves the expected size of the maximum matching. Empirically, we demonstrate the robustness of our approach using the New York City ride-hailing datasets and adversarial synthetic benchmarks. Our results show that near-optimal global matching is achievable even with highly constrained local budgets, significantly outperforming standard online baselines.", "date": "2026-05-13", "categories": []} {"id": "2605.14416", "title": "A Unified Knowledge Embedded Reinforcement Learning-based Framework for Generalized Capacitated Vehicle Routing Problems", "abstract": "The Capacitated Vehicle Routing Problem (CVRP) is a fundamental NP-hard problem with broad applications in logistics and transportation. Real-world CVRPs often involve diverse objectives and complex constraints, such as time windows or backhaul requirements, motivating the development of a unified solution framework. Recent reinforcement learning (RL) approaches have shown promise in combinatorial optimization, yet they rely on end-to-end learning and lack explicit problem-solving knowledge, limiting solution quality. In this paper, we propose a knowledge-embedded framework inspired by the Route-First Cluster-Second heuristics. It incorporates knowledge at two levels: (1) decomposing CVRPs into the route-first and cluster-second subproblems, and (2) leveraging dynamic programming to solve the second subproblem, whose results guide the RL-based constructive solver to solve the first problem. To mitigate partial observability caused by problem decomposition, we introduce a unified history-enhanced context processing module. Extensive experiments show that this framework achieves superior solution quality compared with state-of-the-art learning-based methods, with a smaller gap to classical heuristics, demonstrating strong generalization across diverse CVRP variants.", "date": "2026-05-14", "categories": []} {"id": "2605.14494", "title": "Learning Scenario Reduction for Two-Stage Robust Optimization with Discrete Uncertainty", "abstract": "Two-Stage Robust Optimization (2RO) with discrete uncertainty is challenging, often rendering exact solutions prohibitive. Scenario reduction alleviates this issue by selecting a small, representative subset of scenarios to enable tractable computation. However, existing methods are largely problem-agnostic, operating solely on the uncertainty set without consulting the feasible region or recourse structure. In this paper, we introduce PRISE, a problem-driven sequential lookahead heuristic that constructs reduced scenario sets by evaluating the marginal impact of each scenario. While PRISE yields high-quality scenario subsets, each selection step requires solving multiple subproblems, making it computationally expensive at scale. To address this, we propose NeurPRISE, a neural surrogate model built on a GNN-Transformer backbone that encodes the per-scenario structure via graph convolution and captures cross-scenario interactions through attention. NeurPRISE is trained via imitation learning with a gain-aware ranking objective, which distills marginal gain information from PRISE into a learned scoring function for scenario ranking and selection. Extensive results on three 2RO problems show that NeurPRISE consistently achieves competitive regret relative to comprehensive methods, maintains strong calability with varying numbers of scenarios, and delivers 7-200x speedup over PRISE. NeurPRISE also exhibits strong zero-shot generalization, effectively handling instances with larger problem scales (up to 5x), more scenarios (up to 4x), and distribution shifts.", "date": "2026-05-14", "categories": []} {"id": "2605.14662", "title": "Scalable Solution of the Stochastic Multi-path Traveling Salesman Problem via Neural Networks", "abstract": "The multi-path Traveling Salesman Problem with stochastic travel costs arises in hybrid vehicle routing applications designed for Smart City and City Logistics, where multiple paths exist between each pair of locations. Travel times along these paths are typically affected by real-time traffic conditions and therefore modeled as stochastic. The objective of the problem is to determine a Hamiltonian tour that minimizes the expected total travel cost under uncertainty. In this work, we adopt a two-stage stochastic programming formulation. In the first stage, a predefined route specifying the sequence of locations to be visited is determined, while taking into consideration a second-stage recourse problem that selects the optimal path from the feasible set of alternative paths for each pair of locations, once real-time traffic conditions are realized. To reduce the computational burden imposed by the large number of scenarios required to capture travel time uncertainty, the innovation of this work is the integration of neural network-based surrogate models to approximate the expected value of the second-stage recourse problem. Different architectures and training strategies for the neural networks are proposed and analyzed, with performance evaluated in terms of computation time, solution quality, and generalization capability. Preliminary findings demonstrate the enhanced scalability and practical applicability of the approach for complex vehicle routing problems under uncertainty.", "date": "2026-05-14", "categories": []} {"id": "2602.13155", "title": "Learning to Approximate Uniform Facility Location via Graph Neural Networks", "abstract": "Neural networks, particularly message-passing neural networks (MPNNs), are increasingly used as heuristics for hard combinatorial optimization problems. Yet many learning-based methods rely on supervision, reinforcement learning, or gradient estimators, causing high computational cost, unstable training, or limited guarantees. Classical approximation algorithms provide worst-case guarantees but are non-differentiable and cannot adapt to structure in natural input distributions. We study this tradeoff through Uniform Facility Location (UniFL), a problem with applications in clustering, summarization, logistics, and supply chains. We propose a fully differentiable MPNN that incorporates approximation-algorithmic principles without solver supervision or discrete relaxations. The model has provable approximation guarantees and empirically improves on standard approximation algorithms, narrowing the gap to integer linear programming.", "date": "2026-02-13", "categories": []} {"id": "2602.17346", "title": "Partial Optimality in the Preordering Problem", "abstract": "Preordering is a generalization of clustering and partial ordering with applications in bioinformatics and social network analysis. Given a finite set $V$ and a value $c_{ab} \\in \\mathbb{R}$ for every ordered pair $ab$ of elements of $V$, the preordering problem asks for a preorder $\\lesssim$ on $V$ that maximizes the sum of the values of those pairs $ab$ for which $a \\lesssim b$. Building on the state of the art in solving this NP-hard problem partially, we contribute new partial optimality conditions and efficient algorithms for deciding these conditions. In experiments with real and synthetic data, these new conditions increase, in particular, the fraction of pairs $ab$ for which it is decided efficiently that $a \\not\\lesssim b$ in an optimal preorder.", "date": "2026-02-19", "categories": []} {"id": "2605.13072", "title": "Neural QAOA$^{2}$: Differentiable Joint Graph Partitioning and Parameter Initialization for Quantum Combinatorial Optimization", "abstract": "The quantum approximate optimization algorithm (QAOA) holds promise for combinatorial optimization but is constrained by limited qubits. While divide-and-conquer frameworks like QAOA$^{2}$ address scalability by partitioning graphs into subgraphs, existing methods suffer from two fundamental limitations: i) misalignment between heuristic partitioning metrics and quantum optimization goals, and ii) topology-blind parameter initialization that leads to optimization cold starts. To bridge these gaps, we propose Neural QAOA$^{2}$, an end-to-end differentiable framework that jointly generates graph partitions and initial parameters. By integrating a generative evaluative network (GEN), our method utilizes a differentiable quantum evaluator as a high-fidelity performance surrogate to provide direct gradient guidance, enabling the joint generator to learn the intrinsic mapping from graph topology to high-quality partition and parameter configurations. Extensive experiments on 183 QUBO, Ising, and MaxCut instances (21 to 1000 variables) demonstrate that our gradient-driven approach broadly outperforms heuristic baselines, ranking first on 101 instances. It exhibits zero-shot generalization across out-of-distribution graph topologies and scales.", "date": "2026-05-13", "categories": []} {"id": "2605.13673", "title": "Graph Neural Networks with Triangle-Based Messages for the Multicut Problem", "abstract": "The multicut problem is an NP-hard combinatorial optimization problem with diverse applications in fields such as bioinformatics, data mining and computer vision. Graph neural networks have been defined for the multicut problem but can be adapted further to its specific objective function and constraints. In this article, we introduce such an adapted graph neural network architecture in which features are assigned only to edges, and the computation of messages is based on triangles in the underlying graph. Experiments with synthetic and real-world instances with up to 200 nodes show that our method outperforms state-of-the-art heuristic solvers in terms of solution quality while maintaining feasible runtimes. For some instances, our method finds optimal solutions in seconds whereas exact solvers need hours to find and certify optimal solutions.", "date": "2026-05-13", "categories": []} {"id": "2411.19517", "title": "RL-SPH: Learning to Achieve Feasible Solutions for Integer Linear Programs", "abstract": "Primal heuristics play a crucial role in quickly finding feasible solutions for NP-hard integer linear programming (ILP). Although $\\textit{end-to-end learning}$-based primal heuristics (E2EPH) have recently been proposed, they are typically unable to independently generate feasible solutions. To address this challenge, we propose RL-SPH, a novel reinforcement learning-based start primal heuristic capable of independently generating feasible solutions, even for ILP involving non-binary integers. Empirically, RL-SPH rapidly obtains high-quality feasible solutions with a 100% feasibility rate, achieving on average a 28.6$\\times$ lower primal gap and a 2.6$\\times$ lower primal integral compared to existing start primal heuristics.", "date": "2024-11-29", "categories": []} {"id": "2603.24033", "title": "SRG: Score-based Relaxation-guided Generation for Mixed Integer Linear Programming", "abstract": "We propose Score-based Relaxation-guided Generation (SRG), a generative framework based on an approximate formulation of relaxation-guided stochastic differential equations (SDEs) for mixed-integer linear programming. SRG employs a Transformer-based score network that incorporates feasibility and optimality signals into score modeling, encouraging the learned generative model to place more probability mass on feasible, high-quality regions of the solution space. At inference time, SRG directly samples diverse candidate solutions from the learned score model without requiring any additional guidance module. These candidates are then used to construct compact trust-region subproblems for standard MILP solvers. Across multiple public benchmarks, SRG matches or improves upon the solution quality of the strongest learning-based baselines, with particularly strong gains in challenging candidate-generation settings. Moreover, SRG shows promising zero-shot transferability to unseen cross-scale and cross-problem instances, improving solver objectives and reducing search time in several cases through higher-quality initial candidates and compact trust-region search.", "date": "2026-03-25", "categories": []} {"id": "2605.07744", "title": "Alternating Target-Path Planning for Scalable Multi-Agent Coordination", "abstract": "The concurrent target assignment and pathfinding (TAPF) problem extends multi-agent pathfinding (MAPF) by asking planners to allocate distinct targets and collision-free paths to agents. Prior work on TAPF has relied exclusively on Conflict-Based Search (CBS), which tightly couples target assignment and pathfinding, resulting in compute-intensive, non-scalable solutions. In contrast, we propose an iterative refinement framework that decouples target assignment from pathfinding. Our framework builds on modern, fast, suboptimal MAPF solvers, such as LaCAM. Specifically, within a given time budget, it repeatedly solves MAPF for the current target assignment, identifies bottleneck agents via MAPF feedback, and refines the assignment. Empirical results show that feedback-driven reassignment loop is effective, enabling our framework to scale well beyond the reach of the state-of-the-art CBS-based solver while maintaining decent solution quality. This represents a solid step toward practical, large scale TAPF suitable for real-world setups.", "date": "2026-05-08", "categories": []} {"id": "2605.09277", "title": "Worst-Case Regret Bounds for Combinatorial Thompson Sampling in Sleeping Semi-Bandits", "abstract": "We revisit combinatorial Thompson sampling (CTS) for semi-bandits with sleeping arms, where arm availability varies over time and actions must satisfy combinatorial constraints, as in wireless mesh routing with fluctuating link availability. Despite its practical relevance, CTS has been hindered by several long-standing problems: (i) the absence of worst-case regret guarantees in the semi-bandit setting even without sleeping arms, (ii) the lack of theory under adversarially varying availability, and (iii) the consistently weak empirical performance of CTS with Gaussian priors (CTS-G). This paper resolves these long-standing issues by providing the first worst-case regret analysis of CTS-G, proving an upper bound of $\\tilde{O}(m\\sqrt{NT})$ and a matching lower bound of $\\tilde{\\Omega}(m\\sqrt{NT})$. To bridge the gap between theory and practice, we further propose CL-SG, a simple CTS-G variant that samples a single shared Gaussian seed each round to coordinate exploration across arms. We show that CL-SG achieves an improved regret bound of $\\tilde{O}(\\sqrt{mNT})$, together with a matching lower bound $\\Omega(\\sqrt{mNT})$. Experiments on real-world datasets demonstrate that CL-SG consistently outperforms strong baselines including CTS-G and CTS-B, and we open-source our implementation for reproducibility.", "date": "2026-05-10", "categories": []} {"id": "2605.11910", "title": "Rethinking Positional Encoding for Neural Vehicle Routing", "abstract": "Transformer-based models have become the dominant paradigm for neural combinatorial optimization (NCO) of vehicle routing problems (VRPs), yet the role of positional encoding (PE) in these architectures remains largely unexplored. Unlike natural language, where tokens are uniformly spaced on a line, routing solutions exhibit several properties that render standard NLP positional encodings inadequate. In this work, we formalize three such structural properties that a routing-aware PE should respect, namely anisometric node distances, cyclic and direction-aware topology, and hierarchical depot-anchored global multi-route structure, combining them with a unifying design principle of geometric grounding. Guided by these criteria, we analyze and compare PE methods spanning NLP, graph-transformer, and routing-specific families, and propose a hierarchical anisometric PE that combines a distance-indexed, circularly consistent in-route encoding with a depot-anchored angular cross-route encoding. Extensive experiments across diverse VRP variants demonstrate that geometry-grounded PE consistently outperforms index-based alternatives, with gains that transfer across problem variants, model architectures, and distribution shifts.", "date": "2026-05-12", "categories": []} {"id": "2605.12365", "title": "QAP-Router: Tackling Qubit Routing as Dynamic Quadratic Assignment with Reinforcement Learning", "abstract": "Qubit routing is a fundamental problem in quantum compilation, known to be NP-hard. Its dynamic nature makes local routing decisions propagate and compound over time, making global efficient solutions challenging. Existing heuristic methods rely on local rules with limited lookahead, while recent learning-based approaches often treat routing as a generic sequential decision problem without fully exploiting its underlying structure. In this paper, we introduce QAP-Router, framing qubit routing based on a dynamic Quadratic Assignment Problem (QAP) formulation. By modeling logical interactions, or quantum gates, as flow matrices and hardware topology as a distance matrix, our approach captures the interaction-distance coupling in a unified objective, which defines the reward in the reinforcement learning environment. To further exploit this structure, the policy network employs a solution-aware Transformer backbone that encodes the interaction between the flow matrix and the distance matrix into the attention mechanism. We also integrate a lookahead mechanism that blends naturally into the QAP framework, preventing myopic decisions. Extensive experiments on 1,831 real-world quantum circuits from the MQTBench, AgentQ and QUEKO datasets show that our method substantially reduces the CNOT gate count of routed circuits by 15.7%, 30.4% and 12.1%, respectively, relative to existing industry compilers.", "date": "2026-05-12", "categories": []} {"id": "2510.19544", "title": "Demonstrating Real Advantage of Machine-Learning-Enhanced Monte Carlo for Combinatorial Optimization", "abstract": "Combinatorial optimization problems are central to both practical applications and the development of optimization methods. While classical and quantum algorithms have been refined over decades, machine learning--assisted approaches are comparatively recent and have not yet consistently outperformed simple, state-of-the-art classical methods. Here, we focus on a class of Quadratic Unconstrained Binary Optimization (QUBO) problems, specifically the challenge of finding minimum energy configurations in three-dimensional Ising spin glasses. We use a Global Annealing Monte Carlo algorithm that integrates standard local moves with global moves proposed via machine learning. We show that local moves play a crucial role in achieving optimal performance. Benchmarking against Simulated Annealing and Population Annealing, we demonstrate that Global Annealing not only surpasses the performance of Simulated Annealing but also exhibits greater robustness than Population Annealing, maintaining effectiveness across problem hardness and system size without hyperparameter tuning. These results provide clear and robust evidence that a machine learning--assisted optimization method can exceed the capabilities of classical state-of-the-art techniques in a combinatorial optimization setting.", "date": "2025-10-22", "categories": []} {"id": "2602.06733", "title": "Pairwise is Not Enough: Hypergraph Neural Networks for Multi-Agent Pathfinding", "abstract": "Multi-Agent Path Finding (MAPF) is a representative multi-agent coordination problem, where multiple agents are required to navigate to their respective goals without collisions. Solving MAPF optimally is known to be NP-hard, leading to the adoption of learning-based approaches to alleviate the online computational burden. Prevailing approaches, such as Graph Neural Networks (GNNs), are typically constrained to pairwise message passing between agents. However, this limitation leads to suboptimal behaviours and critical issues, such as attention dilution, particularly in dense environments where group (i.e. beyond just two agents) coordination is most critical. Despite the importance of such higher-order interactions, existing approaches have not been able to fully explore them. To address this representational bottleneck, we introduce HMAGAT (Hypergraph Multi-Agent Attention Network), a novel architecture that leverages attentional mechanisms over directed hypergraphs to explicitly capture group dynamics. Empirically, HMAGAT establishes a new state-of-the-art among learning-based MAPF solvers: e.g., despite having just 1M parameters and being trained on 100$\\times$ less data, it outperforms the current SoTA 85M parameter model. Through detailed analysis of HMAGAT's attention values, we demonstrate how hypergraph representations mitigate the attention dilution inherent in GNNs and capture complex interactions where pairwise methods fail. Our results illustrate that appropriate inductive biases are often more critical than the training data size or sheer parameter count for multi-agent problems.", "date": "2026-02-06", "categories": []} {"id": "2602.22551", "title": "Identifying Multi-Hit Cancer Drivers Without Massive Parallelization: A CP, MIP, and Column Generation Framework", "abstract": "Cancer is often driven by specific combinations of an estimated two to nine gene mutations, known as multi-hit combinations. Identifying these multi-hit combinations of gene mutations that drive cancer is critical for understanding carcinogenesis and designing targeted therapies. We formalize this challenge as the Multi-Hit Cancer Driver Set Cover Problem (MHCDSCP), optimizing the selection of gene combinations to maximize tumor coverage while strictly minimizing normal sample misclassification. While existing approaches rely on exhaustive enumeration and massive parallelization, we introduce fast heuristics based on constraint programming and mixed integer programming formulations. Evaluated on real-world cancer genomics data, our framework matches state-of-the-art supercomputing methods using a single commodity CPU in under a minute. We also propose a price-and-branch heuristic which, by solving the root node to optimality, provides the first provably optimal solutions for over half of the benchmark instances, thereby verifying the near-optimality of our fast heuristics. These findings demonstrate that on real-world problem instances, the MHCDSCP is far less computationally demanding than previously believed, providing an accessible baseline that enables the exploration of previously intractable multi-hit modeling assumptions.", "date": "2026-02-26", "categories": []} {"id": "2604.11101", "title": "Generating Hadamard matrices with transformers", "abstract": "We present a new method for constructing Hadamard matrices that combines transformer neural networks with local search in the PatternBoost framework. Our approach is designed for extremely sparse combinatorial search problems and is particularly effective for Hadamard matrices of Goethals--Seidel type, where Fourier methods permit fast scoring and optimisation. For orders between 100 and 200, it produces large numbers of inequivalent Hadamard matrices, and for larger orders, it succeeds where local search from random initialisation fails. The largest example found by our method has order 252. In addition to these new constructions, our experiments reveal that the transformer can discover and exploit useful hidden symmetry in the search space.", "date": "2026-04-13", "categories": []} {"id": "2605.08756", "title": "AHD Agent: Agentic Reinforcement Learning for Automatic Heuristic Design", "abstract": "Automatic heuristic design (AHD) has emerged as a promising paradigm for solving NP-hard combinatorial optimization problems (COPs). Recent works show that large language models (LLMs), when integrated into well-designed frameworks (i.e., LLM-AHD), can autonomously discover high-performing heuristics. However, existing LLM-AHD frameworks typically treat LLMs as passive generators within fixed workflows, where the model generates heuristics from manually designed, limited context. Such context may fail to capture state-dependent information (e.g., specific failure modes), leading to inefficient trial-and-error exploration. To overcome these limitations, we propose AHD Agent, a novel tool-integrated, multi-turn framework that empowers LLMs to proactively decide whether to generate heuristics or invoke tools to retrieve targeted evidence from the solving environment. To effectively train such a dynamic decision-making agent, we introduce an agentic reinforcement learning (RL) system, which leverages a novel environment synthesis pipeline to optimize a compact model's generalizable AHD capabilities. Experiments across eight diverse domains, including four held-out tasks, demonstrate that our 4B-parameter agent matches or surpasses state-of-the-art baselines using much larger models, while requiring significantly fewer evaluations. Model and inference scaling analysis further reveals that AHD Agent offers an effective trajectory toward truly autonomous heuristic design.", "date": "2026-05-09", "categories": []} {"id": "2605.09301", "title": "Neural Cluster First, Route Second: One-Shot Capacitated Vehicle Routing via Differentiable Optimal Transport", "abstract": "The Capacitated Vehicle Routing Problem (CVRP) underpins modern last-mile logistics. Current Neural Combinatorial Optimization (NCO) methods construct CVRP solutions autoregressively, inheriting sequential decoding bottlenecks, sensitivity to spatial symmetries, and brittle out-of-distribution behavior. We revisit the classical Cluster-First-Route-Second (CFRS) paradigm -- long known to be asymptotically optimal but largely overlooked by NCO -- and argue that it is structurally aligned with the core strengths of deep learning: similarity and assignment over global context, rather than the construction of long sequential tours. We introduce Neural CFRS, the first purely non-autoregressive one-shot neural CFRS framework for the CVRP. It enforces global fleet-capacity constraints end-to-end via a differentiable entropic Optimal Transport layer, producing a continuous transport plan to sparsify an exact capacitated assignment solver. We provide formal theoretical guarantees that our architecture intrinsically abstracts away $E(2)$ spatial, inter-route permutation, and intra-route traversal symmetries. By equipping the framework with a pre-trained spatial vocabulary, we unlock extreme parameter efficiency and zero-shot scaling. Designed primarily for real-world spatial distributions under a constant capacity setting, Neural CFRS scales robustly to out-of-distribution $N=1000$ instances with a < 4% gap -- retaining an approximate 5% gap at this scale even as an ultra-lightweight, single-layer architecture. Furthermore, when deployed out-of-the-box on standard benchmarks, we achieve a highly competitive 2.73% optimality gap on size-100 problems.", "date": "2026-05-10", "categories": []} {"id": "2605.10122", "title": "Rethinking Constraint Awareness for Efficient State Embedding of Neural Routing Solver", "abstract": "Heavy-Encoder-Light-Decoder (HELD) neural routing solvers have emerged as a promising paradigm due to their broad applicability across multiple vehicle routing problems (VRPs). However, they typically struggle with VRP variants with complex constraints. To address this limitation, this paper systematically revisits existing neural solvers from the perspective of the generation mechanism for state embeddings (i.e., query vector prior to compatibility calculation) during decoding. We identify that current mechanisms restrict the observation space during attention computation, introducing a key bottleneck to achieving high-quality solutions. Through detailed empirical analysis, we demonstrate the necessity of preserving a global observation space. To overcome the constraint-agnostic drawback inherent to global observation spaces, we propose a simple yet powerful Constraint-Aware Residual Modulation (CARM) module. By adaptively modulating the context embedding with constraint-relevant variables, CARM effectively enhances constraint awareness, enabling the neural solver to fully leverage the global observation space and generate an efficient state embedding. Extensive experimental results across two single-task and five multi-task neural routing solvers confirm that the CARM module consistently boosts baseline performance. Notably, solvers equipped with our CARM achieve substantial improvements in scaling to large-scale instances and in generalizing to unseen VRP variants. These findings provide valuable insights for the architectural design of neural routing solvers.", "date": "2026-05-11", "categories": []} {"id": "2605.10257", "title": "Towards Autonomous Railway Operations: A Semi-Hierarchical Deep Reinforcement Learning Approach to the Vehicle Rescheduling Problem", "abstract": "Managing disruptions in railway traffic management is a major challenge. Rising traffic density and infrastructure limits increase complexity, making the Vehicle Routing and Scheduling Problem (VRSP) difficult to solve reliably and in real time. While Operational Research (OR) methods are widely used, most dispatching still relies on human expertise due to the problem's exponential combinatorial complexity. Reinforcement Learning (RL) has gained attention for its potential in multi-agent coordination, but existing RL approaches often underperform OR methods and struggle to scale in dense rail networks. This paper addresses this gap from a machine learning perspective by introducing a semi-hierarchical RL formulation tailored to operational railway constraints. The method separates dispatching from routing through dedicated action and observation spaces, enabling policies to specialise in distinct decision scopes and addressing the imbalance between rare dispatch decisions and frequent routing updates. The approach is evaluated on the Flatland-RL simulator across five difficulty levels and 50 random seeds, with 7 to 80 trains. Results show substantially improved coordination, resource utilisation, and robustness compared with heuristic baselines and monolithic RL, nearly doubling the number of trains reaching their destinations, while keeping deadlock rates below 5% and adaptively sequencing, delaying, or cancelling trains under heavy congestion.", "date": "2026-05-11", "categories": []} {"id": "2605.10917", "title": "Optimal and Scalable MAPF via Multi-Marginal Optimal Transport and Schr\\\"odinger Bridges", "abstract": "We consider anonymous multi-agent path finding (MAPF) where a set of robots is tasked to travel to a set of targets on a finite, connected graph. We show that MAPF can be cast as a special class of multi-marginal optimal transport (MMOT) problems with an underlying Markovian structure, under which the exponentially large MMOT collapses to a linear program (LP) polynomial in size. Focusing on the anonymous setting, we establish conditions under which the corresponding LP is feasible, totally unimodular, and consequently, yields min-cost, integral $(\\{0,1\\})$ transports that do not overlap in both space and time. To adapt the approach to large-scale problems, we cast the MAPF-MMOT in a probabilistic framework via Schr\\\"odinger bridges. Under standard assumptions, we show that the Schr\\\"odinger bridge formulation reduces to an entropic regularization of the corresponding MMOT that admits an iterative Sinkhorn-type solution. The Schr\\\"odinger bridge, being a probabilistic framework, provides a shadow (fractional) transport that we use as a template to solve a reduced LP and demonstrate that it results in near-optimal, integral transports at a significant reduction in complexity. Extensive experiments highlight the optimality and scalability of the proposed approaches.", "date": "2026-05-11", "categories": []} {"id": "2503.12285", "title": "A Resilience Framework for Bi-Criteria Combinatorial Optimization with Bandit Feedback", "abstract": "We study bi-criteria combinatorial optimization under noisy function evaluations. While resilience and black-box offline-to-online reductions have been studied in single-objective settings, extending these ideas to bi-criteria problems introduces new challenges due to the coupled degradation of approximation guarantees for objectives and constraints. We introduce a notion of $(\\alpha,\\beta,\\delta,\\texttt{N})$-resilience for bi-criteria approximation algorithms, capturing how joint approximation guarantees degrade under bounded (possibly worst-case) oracle noise, and develop a general black-box framework that converts any resilient offline algorithm into an online algorithm for bi-criteria combinatorial multi-armed bandits with bandit feedback. The resulting online guarantees achieve sublinear regret and cumulative constraint violation of order $\\tilde{O}(\\delta^{2/3}\\texttt{N}^{1/3}T^{2/3})$ without requiring structural assumptions such as linearity, submodularity, or semi-bandit feedback on the noisy functions. We demonstrate the applicability of the framework by establishing resilience for several classical greedy algorithms in submodular optimization.", "date": "2025-03-15", "categories": []} {"id": "2605.06682", "title": "Fast and Effective Redistricting Optimization via Composite-Move Tabu Search", "abstract": "Spatial redistricting is a practical combinatorial optimization problem that demands high-quality solutions, rapid turnaround, and flexibility to accommodate multi-criteria objectives and interactive refinement. A central challenge is the contiguity constraint: enforcing contiguity in integer-programming or heuristic search can severely shrink the feasible neighborhood, weaken exploration, and trap the search in poor local optima. We introduce a composite-move Tabu search (CM-Tabu) that systematically expands the feasible neighborhood space in Tabu search while preserving contiguity. When a boundary unit cannot be reassigned individually without disconnecting its district, our method identifies a minimal set of units that can move together, or a pair of units (or sets of units) that can be switched, as a contiguity-preserving composite move. Candidate single-unit and composite moves are generated in linear time by analyzing each district's contiguity graph using articulation points and biconnected components. Extensive experiments demonstrate that the proposed approach substantially improves solution quality, run-to-run robustness, and computational efficiency relative to traditional Tabu search and other baselines. For example, in the Philadelphia case, the approach can consistently attain the theoretical global optimum in population-equality and support multi-criteria trade-offs. CM-Tabu delivers optimization performance suitable for real-world practices and decision-support workflows.", "date": "2026-04-24", "categories": []} {"id": "2605.06900", "title": "Accelerated Relax-and-Round for Concave Coverage Problems", "abstract": "We present an accelerated relax-and-round algorithm for concave coverage problems, which generalize the classic maximum coverage problem. Building on the relax-and-round framework of Barman et al. [STACS 2021], we propose two significant improvements. First, we replace the linear programming (LP) relaxation step with a projected accelerated gradient method applied to a smooth surrogate objective to achieve a $\\widetilde{O}(mn \\varepsilon^{-1})$ running time. Second, we use a specialized rounding scheme for the hypersimplex that combines the Carath\\'eodory decomposition algorithm in Karalias et al. [NeurIPS 2025] with randomized swap rounding of Chekuri et al. [FOCS 2010]. We prove tight approximation ratios for new reward functions, including a $0.827$-approximation for the logarithmic reward $\\varphi(x) = \\log(1 + x)$. Finally, we conduct maximum multi-coverage experiments on synthetic and real-world graphs, demonstrating that our algorithm outperforms approaches that use state-of-the-art LP solvers.", "date": "2026-05-07", "categories": []} {"id": "2605.06931", "title": "Target-Aware Data Augmentation for SAT Prediction", "abstract": "Learning-based approaches to NP-hard problems have shown increasing promise, but their progress is fundamentally constrained by the high cost of generating labeled training data. In domains such as Boolean satisfiability (SAT), standard pipelines rely on solver-in-the-loop labeling, which scales poorly with problem size and limits the amount of usable supervision. This bottleneck hinders the broader goal of leveraging machine learning to capture structure in hard combinatorial problems. In this work, we propose a target-aware, solver-free data generation framework for SAT that produces correctly labeled SAT and UNSAT instances by construction, eliminating the need for expensive solver calls. Our method aligns generated instances with the structural properties of a target benchmark, making synthetic data effective for downstream learning. We further develop a linear-programming-aware graph neural network (LPGNN) architecture that incorporates constraint-violation residuals into message passing, enabling the model to exploit underlying optimization structure. Together, these contributions support a data-centric paradigm for learning on NP-hard problems, where scalable, task-aligned data generation is as critical as model design. Our approach yields orders-of-magnitude speedups in data generation, demonstrating that benchmark-aligned synthetic data can effectively augment solver-labeled datasets for GNN-based SAT prediction.", "date": "2026-05-07", "categories": []} {"id": "2605.07080", "title": "Online Allocation with Unknown Shared Supply", "abstract": "Many real-world resource allocation systems, such as humanitarian logistics and vaccine distribution, must preposition limited supply across multiple locations before demand is realized while stockouts incur irreversible service losses. To study this, we introduce the Online Shared Supply Allocation (OSSA) problem, a stateful online model in which a central hub allocates a finite, unknown supply to multiple sites facing sequential demand under fixed-charge transportation costs and lost-sales penalties. Unlike classical make-to-stock or make-to-order inventory models, OSSA precludes backlogging and replenishment only hedges against future demand. To tackle OSSA, we propose a deterministic threshold-proportional policy GPA and prove that it achieves a $4/3$-approximation to the offline optimum up to an additive term independent of the total supply. We complement this with matching lower bounds showing that the $4/3$ ratio is tight and that the additive-error dependence is unavoidable, even for randomized algorithms that know the total supply upfront. Finally, we develop a learning-augmented extension to GPA that principally incorporates imperfect forecasts (e.g., from human experts or ML models) commonly available in practice, enabling us to exploit high-quality advice while being robust against arbitrary bad ones. Synthetic and real-world experiments show that GPA outperforms natural baselines with global supply is scarce.", "date": "2026-05-08", "categories": []} {"id": "2605.07113", "title": "Solving Max-Cut to Global Optimality via Feasibility-Preserving Graph Neural Networks", "abstract": "Exact solution of hard combinatorial optimization problems often relies on strong convex relaxations, but solving these relaxations repeatedly inside a branch-and-bound algorithm can be prohibitively expensive. Hence, we consider this challenge for Max-Cut, where branch and bound commonly uses semidefinite programming (SDP) relaxations to bound subproblems. We propose a Max-Cut-specific graph neural network that serves as a principled, lightweight neural proxy for these SDP solvers and can be plugged directly into an exact branch-and-bound framework. The proposed architecture has update steps of complexity $\\mathcal{O}(n^2 + ne)$, and predicts both primal- and dual-feasible SDP solutions. The primal SDP solutions yield feasible Max-Cut solutions via the Goemans--Williamson algorithm. In addition, it is trained in a self-supervised fashion without requiring solved SDP relaxations as labels. Empirically, we show that our architecture can substantially reduce the cost of bounding in exact Max-Cut solving by up to $10.6 \\times$ compared with using the state-of-the-art SDP solver Mosek. Our work highlights the potential of learned, validity-preserving surrogates for accelerating exact optimization over structured convex relaxations.", "date": "2026-05-08", "categories": []} {"id": "2605.07214", "title": "HMACE: Heterogeneous Multi-Agent Collaborative Evolution for Combinatorial Optimization", "abstract": "Large Language Models have recently emerged as a promising paradigm for automated heuristic design for NP-hard combinatorial optimization problems. Despite this progress, existing LLM-based methods typically rely on monolithic workflows constrained by rigid templates, thereby restricting memory-guided exploration and triggering premature convergence to local optima. To design an autonomous and collaborative architecture, we introduce HMACE, a Heterogeneous Multi-Agent Collaborative Evolution framework that reconceptualizes heuristic search as an organizational design problem. HMACE decomposes each evolutionary generation into an autonomous, role-specialized loop with four coordinated agents: a Proposer for strategy exploration, a Generator for executable heuristic synthesis, an Evaluator for empirical assessment, and a Reflector for archive-backed memory update. By coupling behavior-aware retrieval, lightweight candidate filtering, and fitness-grounded archive updates, HMACE guides the search toward diverse and promising heuristic behaviors while avoiding redundant evaluations. Extensive evaluations on representative COPs, including TSP, Online BPP, MKP, and PFSP, show that HMACE achieves a favorable quality-efficiency trade-off compared to state-of-the-art single-agent and multi-agent baselines. In the matched LLM-driven reference comparison, HMACE achieves the lowest average gaps on TSP and Online BPP (0.464\\% and 0.223\\%, respectively), while requiring only 0.13M and 0.42M tokens for the two tasks, substantially fewer than the compared baselines.", "date": "2026-05-08", "categories": []} {"id": "2605.07902", "title": "Curvature Beyond Positivity: Greedy Guarantees for Arbitrary Submodular Functions", "abstract": "Submodular functions -- functions exhibiting diminishing returns -- are central to machine learning. When the objective is monotone and non-negative, the greedy algorithm achieves a tight $63\\%$ approximation. But many practical objectives incorporate costs that make them negative on some inputs, and all existing multiplicative guarantees require non-negativity. Prior work handles negativity through additive bounds for the special class of decomposable functions and non-monotonicity through partial-monotonicity parameters, but these address each difficulty in isolation and neither extends the classical structural theory. We extend \\emph{curvature} -- a parameter measuring how far a function deviates from linearity -- to all submodular functions, handling both non-monotonicity and negativity through a single classical concept. A greedy algorithm with pruning achieves a curvature-controlled multiplicative ratio for \\emph{any} submodular function, including those taking negative values -- the first such guarantee beyond monotonicity and non-negativity. In the non-monotone regime $1 \\le c_g < 2.2$, the bound strictly beats the best known uniform ratio of $0.401$ (for non-negative $f$), and it recovers the classical $(1-e^{-c_g})/c_g$ guarantee for monotone functions. A multilinear-extension variant extends the framework to general combinatorial constraints via multilinear relaxation. Experiments on cost-penalized experimental design, coverage, feature selection, and a curvature sweep on Multi-News passage selection support the theory.", "date": "2026-05-08", "categories": []} {"id": "2411.18954", "title": "ReMAP: Neural Reparameterization for Scalable MAP Inference in Arbitrary-Order Markov Random Fields", "abstract": "Scalable high-quality MAP inference in arbitrary-order Markov Random Fields (MRFs) remains challenging. Approximate message-passing methods are often efficient but can degrade on dense or high-order instances, while exact solvers such as Toulbar2 become increasingly expensive at scale. We present ReMAP, an instance-wise neural reparameterization framework that directly optimizes a differentiable relaxation of the original MRF energy. Instead of relying on supervised labels or amortized training, ReMAP treats each MRF as an independent optimization problem: a Graph Neural Network produces node-wise label distributions, and gradient-based optimization searches for a low-energy discrete solution in an over-parameterized continuous space. The method supports pairwise and arbitrary-order factors, heterogeneous label cardinalities, and efficient GPU execution, without requiring labeled solutions. We show that the relaxed objective is consistent with the discrete MAP problem and analyze how neural over-parameterization can expose low-energy optimization paths unavailable in the original discrete space. Empirically, on synthetic pairwise and high-order MRFs, UAI 2022 inference benchmarks, and real-world Physical Cell Identity (PCI) problems, ReMAP consistently outperforms approximate baselines and often finds lower-energy solutions than Toulbar2 on hard large-scale instances within practical time budgets.", "date": "2024-11-28", "categories": []} {"id": "2512.18034", "title": "Accelerating Discrete Facility Layout Optimization: A Hybrid CDCL and CP-SAT Architecture", "abstract": "Discrete facility layout design involves placing physical entities to minimize handling costs while adhering to strict safety and spatial constraints. This combinatorial problem is typically addressed using Mixed Integer Linear Programming (MILP) or Constraint Programming (CP), though these methods often face scalability challenges as constraint density increases. This study systematically evaluates the potential of Conflict-Driven Clause Learning (CDCL) with VSIDS heuristics as an alternative computational engine for discrete layout problems. Using a unified benchmarking harness, we conducted a controlled comparison of CDCL, CP-SAT, and MILP across varying grid sizes and constraint densities. Experimental results reveal a distinct performance dichotomy: while CDCL struggles with optimization objectives due to cost-blind branching, it demonstrates unrivaled dominance in feasibility detection, solving highly constrained instances orders of magnitude faster than competing paradigms. Leveraging this finding, we developed a novel \"Warm-Start\" hybrid architecture that utilizes CDCL to rapidly generate valid feasibility hints, which are then injected into a CP-SAT optimizer. Our results confirm that this layered approach successfully accelerates exact optimization, using SAT-driven pruning to bridge the gap between rapid satisfiability and proven optimality.", "date": "2025-12-19", "categories": []} {"id": "2601.22509", "title": "Keep Rehearsing and Refining: Lifelong Learning Vehicle Routing under Continually Drifting Tasks", "abstract": "Existing neural solvers for vehicle routing problems (VRPs) are typically trained either in a one-off manner on a fixed set of pre-defined tasks or in a lifelong manner with tasks arriving sequentially, assuming sufficient training on each task. Both settings overlook a common real-world property: problem patterns may drift continually over time, yielding massive tasks sequentially arising, each with only limited training resources. In this paper, we propose a novel lifelong learning paradigm for neural VRP solvers under continual task drift over time, where each task is locally stationary at one learning time step but receives only insufficient training resources. We empirically demonstrate that such continual drift arises in practice using a real-world logistics dataset. We then propose Dual Replay with Experience Enhancement (DREE), a general framework to improve learning efficiency and mitigate catastrophic forgetting under such drift. Extensive experiments based on both the real-world logistics dataset and commonly used synthetic dataset show that, under such continual drift, DREE effectively learns new tasks, preserves prior knowledge, improves generalization to unseen tasks, and can be applied to various existing neural solvers.", "date": "2026-01-30", "categories": []} {"id": "2604.11535", "title": "Problem Reductions at Scale: Agentic Integration of Computationally Hard Problems", "abstract": "Solving an NP-hard optimization problem often requires reformulating it for a specific solver -- quantum hardware, a commercial optimizer, or a domain heuristic. A tool for polynomial-time reductions between hard problems would let practitioners route any supported problem to any supported solver through a single interface. Building such a library at scale, however, has remained out of reach. We show that harness engineering, the practice of designing constraints, verification systems, and feedback loops that channel AI coding agents, can overcome this barrier. Our harness combines a no-code contribution route for domain experts, a multilayer verification stack ranging from type-level checks to agentic feature tests (AI agents role-playing as end users), and a fully automated implementation-review-integration pipeline. In about three months, we built a command-line tool backed by a library of 100+ problem types and 200+ reduction rules in over 170k lines of Rust. The result suggests that a well-engineered harness lets agents build well-tested software at a scale and pace beyond prior reduction-library efforts. Because the reduction graph composes transitively, a new solver registered for any single problem type instantly becomes available to every problem connected by a reduction path. The source code is available at https://github.com/CodingThrust/problem-reductions.", "date": "2026-04-13", "categories": []} {"id": "2605.01120", "title": "New Bounds for Zarankiewicz Numbers via Reinforced LLM Evolutionary Search", "abstract": "The Zarankiewicz number $\\textbf{Z}(m, n, s, t)$ is the maximum number of edges in a bipartite graph $G_{m, n}$ such that there is no complete $K_{s, t}$ bipartite subgraph. We determine for the first time the exact values of three Zarankiewicz numbers: $\\textbf{Z}(11, 21, 3, 3)=116$, $\\textbf{Z}(11, 22, 3, 3)=121$, and $\\textbf{Z}(12, 22, 3, 3)=132$. We further establish lower bounds for 41 more Zarankiewicz numbers, including several that are within one edge of the best known upper bound, and we match the established value in four more closed cases. Our results are obtained using OpenEvolve, an open-source evolutionary algorithm based on Large Language Models (LLMs) that iteratively improves algorithms for generating mathematical constructions by optimizing a reward signal which we tailored for this specific problem. These findings provide new extremal graph constructions and demonstrate the potential of LLM-guided evolutionary search to contribute to mathematical research. In addition to presenting the resulting constructions, we report the generation algorithms produced, describe the relevant implementation details, and provide our computational costs. Our costs are remarkably low, at less than \\$30 for each Zarankiewicz parameter combination, showing that LLM-guided evolutionary search can be an inexpensive, reproducible, and accessible tool for discovering new combinatorial constructions.", "date": "2026-05-01", "categories": []} {"id": "2605.04510", "title": "Predictive and Prescriptive AI toward Optimizing Wildfire Suppression", "abstract": "Intense wildfire seasons require critical prioritization decisions to allocate scarce suppression resources over a dispersed geographical area. This paper develops a predictive and prescriptive approach to jointly optimize crew assignments and wildfire suppression. The problem features a discrete resource-allocation structure with endogenous wildfire demand and non-linear wildfire dynamics. We formulate an integer optimization model with crew assignments on a time-space-rest network, wildfire dynamics on a time-state network, and linking constraints between them. We develop a two-sided branch-and-price-and-cut algorithm based on: (i) a two-sided column generation scheme that generates fire suppression plans and crew routes iteratively; (ii) a new family of cuts exploiting the knapsack structure of the linking constraints; and (iii) novel branching rules to accommodate non-linear wildfire dynamics. We also propose a data-driven double machine learning approach to estimate wildfire spread as a function of covariate information and suppression efforts, mitigating observed confounding between historical crew assignments and wildfire growth. Extensive computational experiments show that the optimization algorithm scales to otherwise intractable real-world instances; and that the methodology can enhance suppression effectiveness in practice, resulting in significant reductions in area burned over a wildfire season and guiding resource sharing across wildfire jurisdictions.", "date": "2026-05-06", "categories": []} {"id": "2605.05330", "title": "Graph Normalization: Fast Binarizing Dynamics for Differentiable MWIS", "abstract": "We introduce Graph Normalization (GN), a principled dynamical system on graphs that serves as a differentiable approximation engine for the NP-hard Maximum Weight Independent Set (MWIS) problem. MWIS encompasses many combinatorial challenges, including optimal assignment, scheduling, set packing, and MAP inference in discrete Markov Random Fields. Unlike Belief Propagation, we prove GN always converges to a binary indicator of a Maximum Independent Set. GN realizes a fast quasi-Newton descent through an exact Majorization-Minimization step, systematically improving the MWIS relaxed primal objective. We establish an equivalence between GN and the Replicator Dynamics of a nonlinear evolutionary game, where vertices compete for inclusion in an independent set. While a non-potential game, the GN game follows Fisher's Fundamental Theorem of Natural Selection, where the average fitness equals the MWIS primal objective and strictly increases. This connection leads to a weighted extension of the Motzkin-Straus theorem, showing MISes are in bijection with the local minima of a quadratic form over a tilted simplex. For the Assignment Problem, GN acts as a variant of the Sinkhorn algorithm that naturally converges to a hard assignment while generalizing to arbitrary constraint graphs. We demonstrate GN's performance as a fast binarization engine for the state-of-the-art Bregman-Sinkhorn relaxed MWIS solver. On real-world benchmarks with up to 1M edges, GN identifies solutions within 1% of the best known results in seconds on a CPU. GN opens new avenues for deep learning architectures requiring differentiable, \"hard\" decisions under constraints, with applications in structured sparse attention, dynamic network pruning, and Mixture-of-Experts. Beyond core AI, the GN framework enables end-to-end learning of constrained optimization in computer vision, computational biology, and resource allocation.", "date": "2026-05-06", "categories": []} {"id": "2605.05389", "title": "Two-Stage Learned Decomposition for Scalable Routing on Multigraphs", "abstract": "Most neural methods for Vehicle Routing Problems (VRPs) are limited to Euclidean settings or simple graphs. In this work, we instead consider multigraphs, where parallel edges represent distinct travel options with varying trade-offs (e.g., distance vs time). Few methods are designed for such formulations and those that do exist face major scalability issues. We mitigate these scalability issues via a Node-Edge Policy Factorization (NEPF) approach, which splits the routing policy into a node permutation stage and an edge selection stage. To enable the decomposition, we introduce a pre-encoding edge aggregation scheme and a non-autoregressive architecture for the edge stage, as well as a hierarchical reinforcement learning method to train the stages jointly. Our experiments across six VRP variants demonstrate that NEPF matches or outperforms the state-of-the-art in terms of solution quality, while being significantly faster in training and inference.", "date": "2026-05-06", "categories": []} {"id": "2605.06332", "title": "LINC: Decoupling Local Consequence Scoring from Hidden Matching in Constructive Neural Routing", "abstract": "Constructive neural routing solvers usually score the next action by matching a decoder context to candidate embeddings, hiding deterministic one-step consequences such as travel, waiting, slack, and capacity changes. We propose LINC (Local Inference via Normed Comparison), a decoder-side candidate decision architecture that computes these consequences explicitly. LINC uses them according to their decision role: centered relative consequences are compared by a shared linear local scorer, while feasible-set summaries modulate the decoder context. This preserves standard global matching and relieves the hidden state from rediscovering transition arithmetic. The Capacitated Vehicle Routing Problem with Time Windows (CVRPTW) serves as the main constrained-routing stress test; the same interface extends to the Capacitated Vehicle Routing Problem (CVRP) and Traveling Salesman Problem (TSP). In particular, for CVRPTW, LINC reduces PolyNet's Solomon/Homberger gaps from 13.83\\%/38.15\\% to 7.26\\%/14.71\\%; for TSP and CVRP, it also improves external-benchmark gaps.", "date": "2026-05-07", "categories": []} {"id": "2605.06516", "title": "Learning to Cut: Reinforcement Learning for Benders Decomposition", "abstract": "Benders decomposition (BD) is a widely used solution approach for solving two-stage stochastic programs arising in real-world decision-making under uncertainty. However, it often suffers from slow convergence as the master problem grows with an increasing number of cuts. In this paper, we propose Reinforcement Learning for BD (RLBD), a framework that adaptively selects cuts using a neural network-based stochastic policy. The policy is trained using a policy gradient method via the REINFORCE algorithm. We evaluate the proposed approach on a two-stage stochastic electric vehicle charging station location problem and compare it with vanilla BD and LearnBD, a supervised learning approach that classifies cuts using a support vector machine. Numerical results demonstrate that RLBD achieves substantial improvements in computational efficiency and exhibits strong generalization to problems with similar structures but varying data inputs and decision variable dimensions.", "date": "2026-05-07", "categories": []} {"id": "2605.06561", "title": "Optimal Counterfactual Search in Tree Ensembles: A Study Across Modeling and Solution Paradigms", "abstract": "Trust in counterfactual explanations depends critically on whether their recommended changes are truly minimal: suboptimal explanations may vastly overshoot the actual changes needed to alter a decision, and heuristic errors can affect individuals unevenly, giving some users relevant recourse while assigning others unnecessarily costly recommendations. Consequently, we study the problem of computing optimal counterfactual explanations for tree ensembles under plausibility and actionability constraints. This is a combinatorial problem: for a fixed model, counterfactual search boils down to selecting consistent branching decisions and threshold-defined regions under a distance objective. We exploit this structure through CPCF, a constraint programming (CP) formulation in which numerical features are encoded as interval domains induced by split thresholds, while discrete features retain native finite-domain representations. This yields a compact finite-domain formulation that supports multiple distance objectives without continuous split-boundary search. We then place CPCF in a broader comparison across mathematical programming paradigms: we extend a maximum Boolean satisfiability (MaxSAT) formulation, originally designed for hard-voting random forests, to soft-voting ensembles, and compare against the current state-of-the-art mixed-integer linear programming (MILP) optimal approach. Across ten datasets and three types of tree ensembles, we analyze scalability, anytime performance, and sensitivity to distance metrics. We observe that CP achieves the best overall performance. More importantly, our results identify regimes in which the specific strengths of each paradigm make it best suited: CP is most versatile overall, MaxSAT handles hard-voting ensembles particularly well, and MILP remains competitive in amortized inference settings with a moderate number of split levels.", "date": "2026-05-07", "categories": []} {"id": "2504.05407", "title": "RouteFormer: A Transformer-Based Routing Framework for Autonomous Vehicles", "abstract": "Autonomous surveillance missions in Internet of Things (IoT) networks often involve solving NP-hard combinatorial optimization problems to ensure efficient resource utilization. To address the limitations of conventional heuristics in dynamic environments, we propose RouteFormer, a novel framework for single-agent routing in graph-based terrains. RouteFormer creates a synergy between the global context awareness of the transformer self-attention mechanism and the adaptive decision-making capabilities of Reinforcement Learning (RL). This architecture allows the system to output optimized routing decisions that adapt to complex task dependencies and resource availability without requiring labeled training datasets. We evaluated RouteFormer on varying graph sizes designed to resemble realistic reconnaissance missions. The results indicate that our model effectively handles the complexity of missions requiring multiple action profiles, outperforming baseline approaches, in terms of both time and distance. Specifically, RouteFormer achieved 10\\% and 7\\% reduction in distance compared to the solutions obtained from well-established solvers like Concorde and Lin-Kernighan-Helsgaun-3 (LKH-3). This improvement was achieved by effectively incorporating mission-specific constraints that traditional solvers overlook. The proposed framework serves as a modular, scalable pipeline for diverse autonomous scheduling and routing tasks.", "date": "2025-04-07", "categories": []} {"id": "2604.16804", "title": "AutoOR: Scalably Post-training LLMs to Autoformalize Operations Research Problems", "abstract": "Optimization problems are central to decision-making in manufacturing, logistics, scheduling, and other industrial settings. Translating complicated descriptions of these problems into solver-ready formulations requires specialized operations research (OR) expertise, making it hard to scale. We present AutoOR, a scalable synthetic data generation and reinforcement learning pipeline that trains LLMs to autoformalize optimization problems specified in natural language across linear, mixed-integer, and non-linear categories. AutoOR generates verified training data from standard optimization forms and uses solver execution feedback as the reward signal for RL post-training. AutoOR applied to an 8B model achieves state-of-the-art or competitive results across six established OR benchmarks, matching significantly larger frontier models. For a non-linear problem class involving physical dynamics, where frontier models score near 0%, we introduce a curriculum RL strategy that bootstraps from limited initial training data to make this class tractable for post-training. We believe that methods such as AutoOR can significantly accelerate industrial decision-making with AI.", "date": "2026-04-18", "categories": []} {"id": "2605.04225", "title": "ARMATA: Auto-Regressive Multi-Agent Task Assignment", "abstract": "Coordinating multi-agent systems over spatially distributed areas requires solving a complex hierarchical problem: first distributing areas among agents (allocation) and subsequently determining the optimal visitation order (routing). Existing methods typically decouple these stages ignoring inter-stage dependencies or rely on decentralized heuristics that lack global context. In this work, we propose a centralized, fully end-to-end auto-regressive framework that jointly generates allocation decisions and routing sequences. The core contribution of our approach is a multi-stage decoding mechanism that unifies high-level allocation and low-level routing in a single autoregressive pass while maintaining a centralized global state. This enables the model to implicitly balance workload distribution with routing efficiency, avoiding local optima common in decentralized methods. Extensive experiments demonstrate that our method significantly outperforms diverse baselines, achieving up to a 20\\% improvement in solution quality over industrial solvers such as Google OR-Tools, IBM CPLEX, and LKH-3, while reducing computation time from hours to seconds.", "date": "2026-05-05", "categories": []} {"id": "2605.03339", "title": "Automated Large-scale CVRP Solver Design via LLM-assisted Flexible MCTS", "abstract": "Solving large-scale CVRP (LSCVRP) with hundreds to thousands of nodes remains difficult for even state-of-the-art solvers. Divide-and-conquer can scale by decomposing the instance into size-reduced subproblems, but designing decomposition logic and configuring sub-solvers is highly expertise- and labor-intensive. Large Language Models (LLMs) have emerged as promising tools for automated algorithm design. However, existing LLM-driven approaches struggle with LSCVRP primarily due to the difficulty in generating sophisticated search strategies within a limited context window. To bridge this gap, we propose the LLM-assisted Flexible Monte Carlo Tree Search (LaF-MCTS), a novel framework that automates the design of high-performance LSCVRP solvers. We develop a three-tier decision hierarchy to enable incremental design of decomposition policies and sub-solvers for LSCVRP. To enable efficient search within the algorithmic hypothesis space, we introduce semantic pruning to eliminate semantically and structurally redundant codes, and branch regrowth to regenerate codes and preserve diversity. Extensive experiments on CVRPLib demonstrate that LaF-MCTS autonomously composes and optimizes decomposition-enhanced solvers that surpasses various state-of-the-art CVRP solvers.", "date": "2026-05-05", "categories": []} {"id": "2605.03842", "title": "SOAR: Real-Time Joint Optimization of Order Allocation and Robot Scheduling in Robotic Mobile Fulfillment Systems", "abstract": "Robotic Mobile Fulfillment Systems (RMFS) rely on mobile robots for automated inventory transportation, coordinating order allocation and robot scheduling to enhance warehousing efficiency. However, optimizing RMFS is challenging due to strict real-time constraints and the strong coupling of multi-phase decisions. Existing methods either decompose the problem into isolated sub-tasks to guarantee responsiveness at the cost of global optimality, or rely on computationally expensive global optimization models that are unsuitable for dynamic industrial environments. To bridge this gap, we propose SOAR, a unified Deep Reinforcement Learning framework for real-time joint optimization. SOAR transforms order allocation and robot scheduling into a unified process by utilizing soft order allocations as observations. We formulate this as an Event-Driven Markov Decision Process, enabling the agent to perform simultaneous scheduling in response to asynchronous system events. Technically, we employ a Heterogeneous Graph Transformer to encode the warehouse state and integrate phased domain knowledge. Additionally, we incorporate a reward shaping strategy to address sparse feedback in long-horizon tasks. Extensive experiments on synthetic and real-world industrial datasets, in collaboration with Geekplus, demonstrate that SOAR reduces global makespan by 7.5\\% and average order completion time by 15.4\\% with sub-100ms latency. Furthermore, sim-to-real deployment confirms its practical viability and significant performance gains in production environments. The code is available at https://github.com/200815147/SOAR.", "date": "2026-05-05", "categories": []} {"id": "2412.13235", "title": "Logic-Constrained Shortest Paths for Flight Planning", "abstract": "The logic-constrained shortest path problem (LCSPP) combines a one-to-one shortest path problem with satisfiability constraints imposed on the routing graph. This setting arises in flight planning, where air traffic control (ATC) authorities are enforcing a set of traffic flow restrictions (TFRs) on aircraft routes in order to increase safety and throughput. We propose a new branch and bound-based algorithm for the LCSPP. The resulting algorithm has three main degrees of freedom: the node selection rule, the branching rule and the conflict. While node selection and branching rules have been long studied in the MIP and SAT communities, most of them cannot be applied out of the box for the LCSPP. We review the existing literature and develop tailored variants of the most prominent rules. The conflict, the set of variables to which the branching rule is applied, is unique to the LCSPP. We analyze its theoretical impact on the B&B algorithm. In the second part of the paper, we show how to model the flight planning problem with TFRs as an LCSPP and solve it using the branch and bound algorithm. We demonstrate the algorithm's efficiency on a dataset consisting of a global flight graph and a set of around 20000 real TFRs obtained from our industry partner Lufthansa Systems GmbH. We make this dataset publicly available. Finally, we conduct an empirical in-depth analysis of dynamic shortest path algorithms, node selection rules, branching rules and conflicts. Carefully choosing an appropriate combination yields an improvement of an order of magnitude compared to an uninformed choice.", "date": "2024-12-17", "categories": []} {"id": "2605.02461", "title": "Middle-mile logistics through the lens of goal-conditioned reinforcement learning", "abstract": "Middle-mile logistics describes the problem of routing parcels through a network of hubs linked by trucks with finite capacity. We rephrase this as a multi-object goal-conditioned MDP. Our method combines graph neural networks with model-free RL, extracting small feature graphs from the environment state.", "date": "2026-05-04", "categories": []} {"id": "2511.04685", "title": "A hybrid solution approach for the Integrated Healthcare Timetabling Competition 2024", "abstract": "In this work, we present the solution approach for the Integrated Healthcare Timetabling Competition 2024 submitted by Team Twente, which ultimately ranked third among the finalists. Our approach combines mixed-integer programming, constraint programming, and simulated annealing in a 3-phase solution approach based on decomposition into subproblems. In addition to describing our approach and design decisions, we share our insights and, for the first time, lower bounds on the optimal solution values for the benchmark instances. We analyze the results based on solution quality for the competition and an extended runtime Additionally, we investigate the different soft constraints and specific parts of the algorithm. Finally, we highlight open problems and future research directions for further improving the approach.", "date": "2025-10-23", "categories": []} {"id": "2604.06940", "title": "A First Guess is Rarely the Final Answer: Learning to Search in the Traveling Salesperson Problem", "abstract": "Most neural solvers for the Traveling Salesperson Problem (TSP) are trained to output a single solution, even though practitioners rarely stop there: at test time, they routinely spend extra compute on sampling or post-hoc search. This raises a natural question: can the search procedure itself be learned? Neural improvement methods take this perspective by learning a policy that applies local modifications to a candidate solution, accumulating gains over an improvement trajectory. Yet learned improvement for TSP remains comparatively immature, with existing methods still falling short of robust, scalable performance. We argue that a key reason is design mismatch: many approaches reuse state representations, architectural choices, and training recipes inherited from single-solution methods, rather than being built around the mechanics of local search. This mismatch motivates NICO-TSP (Neural Improvement for Combinatorial Optimization): a 2-opt improvement framework for TSP. NICO-TSP represents the current tour with exactly $n$ edge tokens aligned with the neighborhood operator, scores 2-opt moves directly without tour positional encodings, and trains via a two-stage procedure: imitation learning to short-horizon optimal trajectories, followed by critic-free group-based reinforcement learning over longer rollouts. Under compute-matched evaluations that measure improvement as a function of both search steps and wall-clock time, NICO-TSP delivers consistently stronger and markedly more step-efficient improvement than prior learned and heuristic search baselines, generalizes far more reliably to larger out-of-distribution instances, and serves both as a competitive replacement for classical local search and as a powerful test-time refinement module for constructive solvers.", "date": "2026-04-08", "categories": []} {"id": "2605.00201", "title": "Matroid Algorithms Under Size-Sensitive Independence Oracles", "abstract": "The standard oracle model for matroid algorithms assumes that each independence query can be answered in constant time, regardless of the size of the queried set. While this abstraction has underpinned much of the theoretical progress in matroid optimization, it masks the true computational effort required by these algorithms. In particular, for natural and widely studied classes such as graphic matroids, even a single independence query can require work linear in the size of the set, making the constant-time assumption implausible.\n We address this gap by introducing a size-sensitive cost model where the cost of a query $Q$ scales with $|Q|$. Nearly linear-time oracle implementations exist for broad families of matroids, and this refined abstraction therefore captures the true cost of query evaluation while allowing for a more faithful comparison between general matroids and their natural special cases.\n Within this framework we study three fundamental algorithmic tasks: finding a basis of a matroid, approximating its rank, and approximating its partition size. We establish tight results, proving nearly matching upper and lower bounds that show the optimal query cost is (up to logarithmic factors) quadratic in the size of the matroid. On the algorithmic side, our upper bounds are realized by explicit procedures that construct the desired solution. On the complexity side, our lower bounds are unconditional and already hold even for weaker distinguishing formulations of the problems. Finally, for matroids with maximum circuit size at most $c$, we show that the quadratic barrier can be broken, providing an algorithm that calculates the maximum-weight basis with expected query cost $\\mathcal{O}(n^{2-1/c} \\log n)$.", "date": "2026-04-30", "categories": []} {"id": "2605.00572", "title": "Instance-Aware Parameter Configuration in Bilevel Late Acceptance Hill Climbing for the Electric Capacitated Vehicle Routing Problem", "abstract": "Algorithm performance in combinatorial optimization is highly sensitive to parameter settings, while a single globally tuned configuration often fails to exploit the heterogeneity of instances. This limitation is particularly evident in the Electric Capacitated Vehicle Routing Problem, where instances differ in structure, demand patterns, and energy constraints. This paper investigates instance-aware parameter configuration for Bilevel Late Acceptance Hill Climbing, a state-of-the-art metaheuristic for the Electric Capacitated Vehicle Routing Problem. An offline tuning procedure is used to obtain instance-specific parameter labels, which are then mapped from instance features via a regression model to enable parameter prediction for unseen instances prior to execution. Experimental results on the IEEE WCCI 2020 benchmark and its extensions show that the proposed approach achieves an average objective value reduction of $0.28\\%$ across eight held-out test instances relative to a globally tuned configuration. This corresponds to a significant cost reduction in multimillion-dollar transportation operations.", "date": "2026-05-01", "categories": []} {"id": "2604.28018", "title": "Design Structure Matrix Modularization with Large Language Models", "abstract": "Design Structure Matrix (DSM) modularization, the task of partitioning system elements into cohesive modules, is a fundamental combinatorial challenge in engineering design. Traditional methods treat modularization as a pure graph optimization, without access to the engineering context embedded in the system. Building on prior work on LLM-based combinatorial optimization for DSM sequencing, this paper extends the method to modularization across five cases and three backbone LLMs. Our method achieves near-reference quality within 30 iterations without requiring specialized optimization code. Counterintuitively, domain knowledge, beneficial in sequencing, consistently impairs performance on more complex DSMs. We attribute this to semantic misalignment between the LLM's functional priors and the purely structural optimization objective, and propose the semantic-alignment hypothesis as a testable condition governing knowledge effectiveness with LLMs. Ablation studies identify the most effective input representation, objective formulation, and solution pool design for practical deployment. These findings offer practical guidance for deploying LLMs in engineering design optimization.", "date": "2026-04-30", "categories": []} {"id": "2604.28102", "title": "FiLMMeD: Feature-wise Linear Modulation for Cross-Problem Multi-Depot Vehicle Routing", "abstract": "Solving practical multi-depot vehicle routing problems (MDVRP) is a challenging optimization task central to modern logistics, increasingly driven by e-commerce. To address the MDVRP's computational complexity, neural-based combinatorial optimization methods offer a promising scalable alternative to traditional approaches. However, neural-based methods typically rely on rigid architectures and input encodings tailored to specific problem formulations. In real-world settings, heterogeneous constraints create multiple MDVRP variants, limiting the applicability of such models. While multi-task learning (MTL) has begun to accelerate the development of unified neural-based solvers, prior works focus almost exclusively on single-depot VRPs, leaving the MDVRP unaddressed. To bridge this gap, we propose Feature-wise Linear Modulation for Cross-Problem Multi-Depot Vehicle Routing (FiLMMeD), a novel unified neural-based model for 24 different MDVRP variants. We introduce three main contributions: (1) to improve the model's generalization, we augment the standard Transformer encoder with Feature-wise Linear Modulation (FiLM), which dynamically conditions learned internal representations based on the active set of constraints; (2) we provide an initial demonstration of Preference Optimization in the MTL setting, establishing it as a superior alternative to Reinforcement Learning for future MTL works; (3) to mitigate the generalization gap caused by the introduction of multi-depot constraints, we introduce a targeted curriculum learning strategy that progressively exposes the model to increasingly more complex constraint interactions. Extensive experiments on 24 MDVRP variants (including 8 novel formulations) and 16 single-depot VRPs confirm the effectiveness of FiLMMeD, which consistently outperforms state-of-the-art baselines. Our code is available at: https://github.com/AJ-Correa/FiLMMeD/tree/main", "date": "2026-04-30", "categories": []} {"id": "2602.20730", "title": "Rethinking Efficiency in Neural Combinatorial Optimization: Batched Preference Optimization with Mamba", "abstract": "We study efficiency as a first-class objective in Neural Combinatorial Optimization (NCO) and present ECO, an efficient learning framework that combines batched preference optimization with a Mamba backbone. Instead of tightly interleaving every policy update with on-policy rollouts, ECO decouples trajectory generation from gradient updates through two stages: supervised warm-up on pre-computed solutions and iterative Direct Preference Optimization (DPO) on batched candidate sets generated by the current policy. We pair this learning pipeline with a mixed Mamba encoder-decoder that reduces memory growth on long sequences and improves hardware utilization. A local-search-guided bootstrapping strategy is further used during training to widen preference margins and stabilize iterative improvement. Importantly, local search is only used to construct stronger preference pairs during training and is never invoked at inference time. On TSP and CVRP, ECO achieves the strongest overall performance among the compared neural baselines while also delivering clear advantages in memory usage and throughput. We provide additional analysis on memory scaling, throughput, and the contribution of each design component.", "date": "2026-02-24", "categories": []} {"id": "2604.25269", "title": "Online combinatorial optimization with stochastic decision sets and adversarial losses", "abstract": "Most work on sequential learning assumes a fixed set of actions that are available all the time. However, in practice, actions can consist of picking subsets of readings from sensors that may break from time to time, road segments that can be blocked or goods that are out of stock. In this paper we study learning algorithms that are able to deal with stochastic availability of such unreliable composite actions. We propose and analyze algorithms based on the Follow-The-Perturbed-Leader prediction method for several learning settings differing in the feedback provided to the learner. Our algorithms rely on a novel loss estimation technique that we call Counting Asleep Times. We deliver regret bounds for our algorithms for the previously studied full information and (semi-)bandit settings, as well as a natural middle point between the two that we call the restricted information setting. A special consequence of our results is a significant improvement of the best known performance guarantees achieved by an efficient algorithm for the sleeping bandit problem with stochastic availability. Finally, we evaluate our algorithms empirically and show their improvement over the known approaches.", "date": "2026-04-28", "categories": []} {"id": "2406.03099", "title": "Supplementary Materials to Graph Convolutional Branch and Bound", "abstract": "This article explores the integration of deep learning models into combinatorial optimization pipelines, specifically targeting NP-hard problems. Traditional exact algorithms for such problems often rely on heuristic criteria to guide the exploration of feasible solutions. In this work, we propose using neural networks to learn informative heuristics, most notably, an optimality score that estimates a solution's proximity to the optimum. This score is used to evaluate nodes within a branch-and-bound framework, enabling a more efficient traversal of the solution space. Focusing on the Traveling Salesman Problem, we introduce Concorde, a state-of-the-art solver, and present a hybrid approach called Graph Convolutional Branch and Bound, which augments it with a graph convolutional neural network trained with a novel unsupervised training strategy that facilitates generalization to graphs of varying sizes without requiring labeled data. Empirical results demonstrate the effectiveness of the proposed method, showing a significant reduction in the number of explored branch-and-bound nodes and overall computational time. Some of the results concerning the use of the 1-tree relaxation are in the supplementary materials.", "date": "2024-06-05", "categories": []} {"id": "2502.03669", "title": "Unrealized Expectations: Comparing AI Methods vs Classical Algorithms for Maximum Independent Set", "abstract": "AI methods, such as generative models and reinforcement learning, have recently been applied to combinatorial optimization (CO) problems, especially NP-hard ones. This paper compares such GPU-based methods with classical CPU-based methods on the Maximum Independent Set (MIS) problem. Strikingly, even on in-distribution random graphs, leading AI-inspired methods are consistently outperformed by the state-of-the-art classical solver KaMIS running on a single CPU, and some AI-inspired methods frequently fail to surpass even the simplest degree-based greedy heuristic. Even with post-processing techniques like local search, AI-inspired methods still perform worse than CPU-based solvers. To better understand the source of these failures, we introduce a novel analysis, serialization, which reveals that non-backtracking AI-inspired methods, e.g. LTFT (which is based on GFlowNets), end up reasoning similarly to the simplest degree-based greedy, and thus worse than KaMIS. More generally, our findings suggest a need for a rethinking of current approaches in AI for CO, advocating for more rigorous benchmarking and the principled integration of classical heuristics. Additionally, we also find that CPU-based algorithm KaMIS have strong performance on sparse random graphs, which appears to show that the shattering threshold conjecture for large independent sets proposed by Coja-Oghlan & Efthymiou (2015) does not apply for real-life sizes (such as 10^6 nodes).", "date": "2025-02-05", "categories": []} {"id": "2601.11883", "title": "Approximation Algorithm for Constrained $k$-Center Clustering: A Local Search Approach", "abstract": "Clustering is a long-standing research problem and a fundamental tool in AI and data analysis. The traditional k-center problem, a fundamental theoretical challenge in clustering, has a best possible approximation ratio of 2, and any improvement to a ratio of 2 - {\\epsilon} would imply P = NP. In this work, we study the constrained k-center clustering problem, where instance-level cannot-link (CL) and must-link (ML) constraints are incorporated as background knowledge. Although general CL constraints significantly increase the hardness of approximation, previous work has shown that disjoint CL sets permit constant-factor approximations. However, whether local search can achieve such a guarantee in this setting remains an open question. To this end, we propose a novel local search framework based on a transformation to a dominating matching set problem, achieving the best possible approximation ratio of 2. The experimental results on both real-world and synthetic datasets demonstrate that our algorithm outperforms baselines in solution quality.", "date": "2026-01-17", "categories": []} {"id": "2601.21048", "title": "Test-Time Adaptation for Unsupervised Combinatorial Optimization", "abstract": "Unsupervised neural combinatorial optimization (NCO) enables learning powerful solvers without access to ground-truth solutions. Existing approaches fall into two disjoint paradigms: models trained for generalization across instances, and instance-specific models optimized independently at test time. While the former are efficient during inference, they lack effective instance-wise adaptability; the latter are flexible but fail to exploit learned inductive structure and are prone to poor local optima. This motivates the central question of our work: how can we leverage the inductive bias learned through generalization while unlocking the flexibility required for effective instance-wise adaptation? We first identify a challenge in bridging these two paradigms: generalization-focused models often constitute poor warm starts for instance-wise optimization, potentially underperforming even randomly initialized models when fine-tuned at test time. To resolve this incompatibility, we propose TACO, a model-agnostic test-time adaptation framework that unifies and extends the two existing paradigms for unsupervised NCO. TACO applies strategic warm-starting to partially relax trained parameters while preserving inductive bias, enabling rapid and effective unsupervised adaptation. Crucially, compared to naively fine-tuning a trained generalizable model or optimizing an instance-specific model from scratch, TACO achieves better solution quality while incurring negligible additional computational cost. Experiments on the canonical problems of minimum vertex cover, maximum clique, maximum independent set, and max cut demonstrate the effectiveness and robustness of TACO across static, distribution-shifted, and dynamic combinatorial optimization problems, establishing it as a practical bridge between generalizable and instance-specific unsupervised NCO.", "date": "2026-01-28", "categories": []} {"id": "2602.20376", "title": "Exploiting Low-Rank Structure in Max-K-Cut Problems", "abstract": "We approach the Max-3-Cut problem through the lens of maximizing complex-valued quadratic forms and demonstrate that low-rank structure in the objective matrix can be exploited, leading to alternative algorithms to classical semidefinite programming (SDP) relaxations and heuristic techniques. We propose an algorithm for maximizing these quadratic forms over a domain of size $K$ that enumerates and evaluates a set of $O\\left(n^{2r-1}\\right)$ candidate solutions, where $n$ is the dimension of the matrix and $r$ represents the rank of an approximation of the objective. We prove that this candidate set is guaranteed to include the exact maximizer when $K=3$ (corresponding to Max-3-Cut) and the objective is low-rank, and provide approximation guarantees when the objective is a perturbation of a low-rank matrix. This construction results in a family of novel, inherently parallelizable and theoretically-motivated algorithms for Max-3-Cut. Extensive experimental results demonstrate that our approach achieves performance comparable to existing algorithms across a wide range of graphs, while being highly scalable.", "date": "2026-02-23", "categories": []} {"id": "2604.21961", "title": "A general optimization solver based on OP-to-MaxSAT reduction", "abstract": "Optimization problems are fundamental in diverse fields, such as engineering, economics, and scientific computing. However, current algorithms are mostly designed for specific problem types and exhibit limited generality in solving multiple types of optimization problems. To enhance generality, we propose an automated reduction method named OP-to-MaxSAT reduction and a general optimization solver based on OP-to-MaxSAT reduction (GORED). GORED unifies the solving of multiple types of optimization problems by reducing the problems from optimization problems to MaxSAT instances in polynomial time and solving them using the state-of-the-art MaxSAT solver. The generality and solution quality of GORED are validated through experiments on 136 instances across 11 types of optimization problems. Experimental results demonstrate that GORED not only successfully solves a wide range of optimization problems but also yields solutions comparable in quality to those from existing methods, with no statistically significant differences observed. By introducing automated reduction, this work shifts the paradigm of optimization solvers from designing specialized algorithms for each problem type to employing a single algorithm for diverse problems. As a result, advances in this single algorithm can now drive progress in a wide range of optimization problems across various domains.", "date": "2026-04-23", "categories": []} {"id": "2604.21029", "title": "Integrated packing, placement, scheduling, and routing of personalized production: a pharmaceutical Industry 4.0 use-case with a planar transport system", "abstract": "The recent emergence of planar transport systems necessitates re-evaluation of Flexible Manufacturing Systems (FMS) to address the simultaneous scheduling of internal logistics and production operations. By operating on a tile-based planar grid, these systems allow independent movers full two-dimensional freedom, mitigating inefficiencies inherent to traditional sequential lines. This paper applies a planar FMS framework to a real-world use case in the pharmaceutical industry: the automated production of personalized drugs.\n Implementing this system requires solving optimization problems at both tactical and operational levels. The tactical level involves decisions regarding production line layout and the positioning of drug dispensers. A Mixed-Integer Quadratic Programming model is utilized for the packing problem to exploit drug co-occurrence patterns found in historical patient data. Subsequently, we solve the placement problem - a bi-level problem combining an assignment problem with Shortest Hamiltonian paths with neighborhoods - to arrange dispensers in a layout minimizing expected travel distances.\n The operational level is encountered daily, scheduling individual movers to process new orders as quickly as possible. This scheduling problem is formulated using Constraint Programming, modeling movers as reservoir resources to ensure order completeness, complemented by a routing phase using an iterative conflict-resolution mechanism and DAG-based reasoning to convert schedules into conflict-free paths.\n Evaluation using real-world prescription data for 40 drugs shows the framework scales efficiently across several layout topologies for up to 500 orders, with schedules that are highly effective and computationally tractable for daily operations.", "date": "2026-04-22", "categories": []} {"id": "2604.21175", "title": "Graph Neural Network-Informed Predictive Flows for Faster Ford-Fulkerson and PAC-Learnability", "abstract": "We propose a learning-augmented framework for accelerating max-flow computation and image segmentation by integrating Graph Neural Networks (GNNs) with the Ford-Fulkerson algorithm. Rather than predicting initial flows, our method learns edge importance probabilities to guide augmenting path selection. We introduce a Message Passing GNN (MPGNN) that jointly learns node and edge embeddings through coupled updates, capturing both global structure and local flow dynamics such as residual capacity and bottlenecks.\n Given an input image, we propose a method to construct a grid-based flow network with source and sink nodes, extract features, and perform a single GNN inference to assign edge probabilities reflecting their likelihood of belonging to high-capacity cuts. These probabilities are stored in a priority queue and used to guide a modified Ford-Fulkerson procedure, prioritizing augmenting paths via an Edmonds-Karp-style search with bottleneck-aware tie-breaking. This avoids repeated inference over residual graphs while leveraging learned structure throughout optimization.\n We further introduce a bidirectional path construction strategy centered on high-probability edges and provide a theoretical framework relating prediction quality to efficiency via a weighted permutation distance metric. Our method preserves max-flow/min-cut optimality while reducing the number of augmentations in practice. We also outline a hybrid extension combining flow warm-starting with edge-priority prediction, establishing a foundation for learning-guided combinatorial optimization in image segmentation.", "date": "2026-04-23", "categories": []} {"id": "2505.02485", "title": "Integrating Column Generation and Large Neighborhood Search for Bus Driver Scheduling with Complex Break Constraints", "abstract": "The Bus Driver Scheduling Problem (BDSP) is a combinatorial optimization problem with the goal to design shifts to cover prearranged bus tours. The objective takes into account the operational cost as well as the satisfaction of drivers. This problem is heavily constrained due to strict legal rules and collective agreements. The objective of this article is to provide state-of-the-art exact and hybrid solution methods that can provide high-quality solutions for instances of different sizes. This work presents a comprehensive study of both an exact method, Branch and Price (B&P), as well as a Large Neighborhood Search (LNS) framework which uses B&P or Column Generation (CG) for the repair phase to solve the BDSP. It further proposes and evaluates a novel deeper integration of B&P and LNS, storing the generated columns from the LNS subproblems and reusing them for other subproblems, or to find better global solutions. The article presents a detailed analysis of several components of the solution methods and their impact, including general improvements for the B&P subproblem, which is a high-dimensional Resource Constrained Shortest Path Problem (RCSPP), and the components of the LNS. The evaluation shows that our approach provides new state-of-the-art results for instances of all sizes, including exact solutions for small instances, and low gaps to a known lower bound for mid-sized instances. Conclusions: We observe that B&P provides the best results for small instances, while the tight integration of LNS and CG can provide high-quality solutions for larger instances, further improving over LNS which just uses CG as a black box. The proposed methods are general and can also be applied to other rule sets and related optimization problems", "date": "2025-05-05", "categories": []} {"id": "2604.20109", "title": "Learning to Solve the Quadratic Assignment Problem with Warm-Started MCMC Finetuning", "abstract": "The quadratic assignment problem (QAP) is a fundamental NP-hard task that poses significant challenges for both traditional heuristics and modern learning-based solvers. Existing QAP solvers still struggle to achieve consistently competitive performance across structurally diverse real-world instances. To bridge this performance gap, we propose PLMA, an innovative permutation learning framework. PLMA features an efficient warm-started MCMC finetuning procedure to enhance deployment-time performance, leveraging short Markov chains to anchor the adaptation to the promising regions previously explored. For rapid exploration via MCMC over the permutation space, we design an additive energy-based model (EBM) that enables an $O(1)$-time 2-swap Metropolis-Hastings sampling step. Moreover, the neural network used to parameterize the EBM incorporates a scalable and flexible cross-graph attention mechanism to model interactions between facilities and locations in the QAP. Extensive experiments demonstrate that PLMA consistently outperforms state-of-the-art baselines across various benchmarks. In particular, PLMA achieves a near-zero average optimality gap on QAPLIB, exhibits remarkably superior robustness on the notoriously difficult Taixxeyy instances, and also serves as an effective QAP solver in bandwidth minimization.", "date": "2026-04-22", "categories": []} {"id": "2603.09172", "title": "Reinforced Generation of Combinatorial Structures: Ramsey Numbers", "abstract": "We present improved lower bounds for nine classical Ramsey numbers: $\\mathbf{R}(3, 13)$ is increased from $60$ to $61$, $\\mathbf{R}(3, 18)$ from $99$ to $100$, $\\mathbf{R}(4, 13)$ from $138$ to $139$, $\\mathbf{R}(4, 14)$ from $147$ to $148$, $\\mathbf{R}(4, 15)$ from $158$ to $159$, $\\mathbf{R}(4, 16)$ from $170$ to $174$, $\\mathbf{R}(4, 18)$ from $205$ to $209$, $\\mathbf{R}(4, 19)$ from $213$ to $219$, and $\\mathbf{R}(4, 20)$ from $234$ to $237$. These results were achieved using AlphaEvolve, an LLM-based code mutation agent. Beyond these new results, we successfully recovered lower bounds for all Ramsey numbers known to be exact, and matched the best known lower bounds across many other cases. These include bounds for which previous work does not detail the algorithms used. Virtually all known Ramsey lower bounds are derived computationally, with bespoke search algorithms each delivering a handful of results. AlphaEvolve is a single meta-algorithm yielding search algorithms for all of our results.", "date": "2026-03-10", "categories": []} {"id": "2604.18645", "title": "On Solving the Multiple Variable Gapped Longest Common Subsequence Problem", "abstract": "This paper addresses the Variable Gapped Longest Common Subsequence (VGLCS) problem, a generalization of the classical LCS problem involving flexible gap constraints between consecutive solutions' characters. The problem arises in molecular sequence comparison, where structural distance constraints between residues must be respected, and in time-series analysis where events are required to occur within specified temporal delays. We propose a search framework based on the root-based state graph representation, in which the state space comprises a generally large number of rooted state subgraphs. To cope with the resulting combinatorial explosion, an iterative beam search strategy is employed, dynamically maintaining a global pool of promising candidate root nodes, enabling effective control of diversification across iterations. To exploit the search for high-quality solutions, several known heuristics from the LCS literature are utilized into the standalone beam search procedure. To the best of our knowledge, this is the first comprehensive computational study on the VGLCS problem comprising 320 synthetic instances with up to 10 input sequences and up to 500 characters. Experimental results show robustness of the designed approach over the baseline beam search in comparable runtimes.", "date": "2026-04-19", "categories": []} {"id": "2206.14234", "title": "PyEPO: A PyTorch-based End-to-End Predict-then-Optimize Library for Linear and Integer Programming", "abstract": "In deterministic optimization, it is typically assumed that all problem parameters are fixed and known. In practice, however, some parameters may be a priori unknown but can be estimated from contextual information. A typical predict-then-optimize approach separates predictions and optimization into two distinct stages. Recently, end-to-end predict-then-optimize has emerged as an attractive alternative. This work introduces the PyEPO package, a PyTorch-based end-to-end predict-then-optimize library in Python. To the best of our knowledge, PyEPO (pronounced like \\textit{pineapple} with a silent ``n\") is the first such generic tool for linear and integer programming with predicted objective function coefficients. It includes various algorithms such as surrogate decision losses, black-box solvers, and perturbed methods. PyEPO offers a user-friendly interface for defining new optimization problems, applying state-of-the-art algorithms, and using custom neural network architectures. We conducted experiments comparing various methods on problems such as Shortest Path, Multiple Knapsack, and Traveling Salesperson Problem, and discussed empirical insights that may guide future research. PyEPO and its documentation are available at https://github.com/khalil-research/PyEPO.", "date": "2022-06-28", "categories": []} {"id": "2503.04798", "title": "Advancing MAPF Toward the Real World: A Scalable Multi-Agent Realistic Testbed (SMART)", "abstract": "We present Scalable Multi-Agent Realistic Testbed (SMART), a realistic and efficient software tool for evaluating Multi-Agent Path Finding (MAPF) algorithms. MAPF focuses on planning collision-free paths for a group of robots. While state-of-the-art MAPF planners can plan paths for hundreds of robots in seconds, they often rely on simplified robot models, making their real-world performance unclear. Researchers typically lack access to hundreds of physical robots in laboratory settings to evaluate the algorithms. Meanwhile, industrial professionals who lack expertise in MAPF require an easy-to-use simulator to efficiently test and understand the performance of MAPF planners in their specific settings. SMART fills this gap with several advantages: (1) SMART uses physics-engine-based simulators to create realistic simulation environments, accounting for complex real-world factors such as robot kinodynamics and execution uncertainties, (2) SMART uses an execution monitor framework based on the Action Dependency Graph, facilitating seamless integration with various MAPF planners and robot models, and (3) SMART scales to thousands of robots. The code is publicly available at https://github.com/smart-mapf/smart with an online service available at https://smart-mapf.github.io/demo/.", "date": "2025-03-03", "categories": []} {"id": "2604.16581", "title": "NCO4CVRP: Neural Combinatorial Optimization for the Capacitated Vehicle Routing Problem", "abstract": "Neural Combinatorial Optimization (NCO) has emerged as a powerful framework for solving combinatorial optimization problems by integrating deep learning-based models. This work focuses on improving existing inference techniques to enhance solution quality and generalization. Specifically, we modify the Random Re-Construct (RRC) approach of the Light Encoder Heavy Decoder (LEHD) model by incorporating Simulated Annealing (SA). Unlike the conventional RRC, which greedily replaces suboptimal segments, our SA-based modification introduces a probabilistic acceptance mechanism that allows the model to escape local optima and explore a more diverse solution space. Additionally, we enhance the Policy Optimization with Multiple Optima (POMO) approach by integrating Beam Search, enabling systematic exploration of multiple promising solutions while maintaining diversity in the search space. We further investigate different inference strategies, including Softmax Sampling, Greedy, Gumbel-Softmax, and Epsilon-Greedy, analyzing their impact on solution quality. Furthermore, we explore instance augmentation techniques, such as horizontal and vertical flipping and rotation-based augmentations, to improve model generalization across different CVRP instances. Our extensive experiments demonstrate that these modifications significantly reduce the optimality gap across various Capacitated Vehicle Routing Problem (CVRP) benchmarks, with Beam Search and SA-based RRC consistently yielding superior performance. By refining inference techniques and leveraging enhanced search strategies, our work contributes to the broader applicability of NCO models in real-world combinatorial optimization tasks.", "date": "2026-04-17", "categories": []} {"id": "2411.09355", "title": "Prices, Bids, Values: One ML-Powered Combinatorial Auction to Rule Them All", "abstract": "We study the design of iterative combinatorial auctions (ICAs). The main challenge in this domain is that the bundle space grows exponentially in the number of items. To address this, recent work has proposed machine learning (ML)-based preference elicitation algorithms that aim to elicit only the most critical information from bidders to maximize efficiency. However, while the SOTA ML-based algorithms elicit bidders' preferences via value queries, ICAs that are used in practice elicit information via \\emph{demand queries}. In this paper, we introduce a novel ML algorithm that provably makes use of the full information from both value and demand queries, and we show via experiments that combining both query types results in significantly better learning performance in practice. Building on these insights, we present MLHCA, a new ML-powered auction that uses value and demand queries. MLHCA significantly outperforms the previous SOTA, reducing efficiency loss by up to a factor 10, with up to 58\\% fewer queries. Thus, MLHCA achieves large efficiency improvements while also reducing bidders' cognitive load, establishing a new benchmark for both practicability and efficiency. Our code is available at https://github.com/marketdesignresearch/MLHCA.", "date": "2024-11-14", "categories": []} {"id": "2604.15837", "title": "Stein Variational Black-Box Combinatorial Optimization", "abstract": "Combinatorial black-box optimization in high-dimensional settings demands a careful trade-off between exploiting promising regions of the search space and preserving sufficient exploration to identify multiple optima. Although Estimation-of-Distribution Algorithms (EDAs) provide a powerful model-based framework, they often concentrate on a single region of interest, which may result in premature convergence when facing complex or multimodal objective landscapes. In this work, we incorporate the Stein operator to introduce a repulsive mechanism among particles in the parameter space, thereby encouraging the population to disperse and jointly explore several modes of the fitness landscape. Empirical evaluations across diverse benchmark problems show that the proposed method achieves performance competitive with, and in several cases superior to, leading state-of-the-art approaches, particularly on large-scale instances. These findings highlight the potential of Stein variational gradient descent as a promising direction for addressing large, computationally expensive, discrete black-box optimization problems.", "date": "2026-04-17", "categories": []} {"id": "2505.13754", "title": "Unsupervised Learning of Local Updates for Maximum Independent Set in Dynamic Graphs", "abstract": "We present the first unsupervised learning model for Maximum-Independent-Set (MaxIS) in dynamic graphs where edges change over time. Our method combines structural learning from graph neural networks (GNNs) with a learned distributed update mechanism that, given an edge addition or deletion event, modifies nodes' internal memories and infers their MaxIS membership in a single, parallel step. We evaluate our model against a mixed integer programming solver and a breadth of unsupervised and supervised learning models for combinatorial optimization on static graphs. Across dynamic graphs of 200-1,000 nodes, our model achieves approximation ratios that are competitive with the state-of-the-art models while running 1.91-6.70x faster. When generalizing to graphs with 100x more nodes than those used for training, our model produces MaxIS solutions 1.00-1.18x larger than all other unsupervised models, but is outperformed by the state-of-the-art supervised model. These results demonstrate that this novel, unsupervised, update-based learning approach to dynamic combinatorial optimization is a viable alternative to the na\\\"ive reapplication of analogous models for static graphs, leveraging temporal information to improve neural methods for combinatorial optimization.", "date": "2025-05-19", "categories": []} {"id": "2601.20868", "title": "Rethinking LLM-Driven Heuristic Design: Generating Efficient and Specialized Solvers via Dynamics-Aware Optimization", "abstract": "Large Language Models (LLMs) have advanced the field of Combinatorial Optimization through automated heuristic generation. Instead of relying on manual design, this LLM-Driven Heuristic Design (LHD) process leverages LLMs to iteratively generate and refine solvers to achieve high performance. However, existing LHD frameworks face two critical limitations: (1) Endpoint-only evaluation, which ranks solvers solely by final gap to a reference solution, ignoring the convergence process and runtime efficiency; (2) High adaptation costs, where distribution shifts necessitate re-adaptation to generate specialized solvers for heterogeneous instance groups. To address these issues, we propose Dynamics-Aware Solver Heuristics (DASH), a framework that co-optimizes solver search mechanisms and runtime schedules guided by a convergence-aware metric, thereby identifying efficient and high-performance solvers. Furthermore, to mitigate expensive re-adaptation, DASH incorporates Profiled Library Retrieval (PLR), which maintains group-specialized solvers for profile-aware warm starts. These solvers are archived concurrently during evolution, allowing DASH to reuse matched specialists across heterogeneous distributions without restarting adaptation. Experiments on four combinatorial optimization problems demonstrate that DASH improves runtime efficiency by over 4 times while outperforming prior LHD baselines in the overall balance between gap and runtime across diverse problem scales. Furthermore, by enabling profile-aware warm starts, DASH maintains lower gap under distribution shift while reducing LLM adaptation costs by about 90%.", "date": "2026-01-14", "categories": []} {"id": "2604.15202", "title": "Benchmarking Classical Coverage Path Planning Heuristics on Irregular Hexagonal Grids for Maritime Coverage Scenarios", "abstract": "Coverage path planning on irregular hexagonal grids is relevant to maritime surveillance, search and rescue and environmental monitoring, yet classical methods are often compared on small ad hoc examples or on rectangular grids. This paper presents a reproducible benchmark of deterministic single-vehicle coverage path planning heuristics on irregular hexagonal graphs derived from synthetic but maritime-motivated areas of interest. The benchmark contains 10,000 Hamiltonian-feasible instances spanning compact, elongated, and irregular morphologies, 17 heuristics from seven families, and a common evaluation protocol covering Hamiltonian success, complete-coverage success, revisits, path length, heading changes, and CPU latency. Across the released dataset, heuristics with explicit shortest-path reconnection solve the relaxed coverage task reliably but almost never produce zero-revisit tours. Exact Depth-First Search confirms that every released instance is Hamiltonian-feasible. The strongest classical Hamiltonian baseline is a Warnsdorff variant that uses an index-based tie-break together with a terminal-inclusive residual-degree policy, reaching 79.0% Hamiltonian success. The dominant design choice is not tie-breaking alone, but how the residual degree is defined when the endpoint is reserved until the final move. This shows that underreported implementation details can materially affect performance on sparse geometric graphs with bottlenecks. The benchmark is intended as a controlled testbed for heuristic analysis rather than as a claim of operational optimality at fleet scale.", "date": "2026-04-16", "categories": []} {"id": "2510.20169", "title": "Empowering Targeted Neighborhood Search via Hyper Tour for Large-Scale TSP", "abstract": "Traveling Salesman Problem (TSP) is a classic NP-hard problem that has garnered significant attention from both academia and industry. While neural-based methods have shown promise for solving TSPs, they still face challenges in scaling to larger instances, particularly in memory constraints associated with global heatmaps, edge weights, or access matrices, as well as in generating high-quality initial solutions and insufficient global guidance for efficiently navigating vast search spaces. To address these challenges, we propose a Hyper Tour Guided Neighborhood Search (HyperNS) method for large-scale TSP instances. Inspired by the ``clustering first, route second\" strategy, our approach initially divides the TSP instance into clusters using a sparse heatmap graph and abstracts them as supernodes, followed by the generation of a hyper tour to guide both the initialization and optimization processes. This method reduces the search space by focusing on edges relevant to the hyper tour, leading to more efficient and effective optimization. Experimental results on both synthetic and real-world datasets demonstrate that our approach outperforms existing neural-based methods, particularly in handling larger-scale instances, offering a significant reduction in the gap to the optimal solution.", "date": "2025-10-23", "categories": []} {"id": "2602.07924", "title": "Optimized Human-Robot Co-Dispatch Planning for Petro-Site Surveillance under Varying Criticalities", "abstract": "Securing petroleum infrastructure requires balancing autonomous system efficiency with human judgment for threat escalation, a challenge unaddressed by classical facility location models assuming homogeneous resources. This paper formulates the Human-Robot Co-Dispatch Facility Location Problem (HRCD-FLP), a capacitated facility location variant incorporating tiered infrastructure criticality, human-robot supervision ratio constraints, and minimum utilization requirements. We evaluate command center selection across three technology maturity scenarios. Results show transitioning from conservative (1:3 human-robot supervision) to future autonomous operations (1:10) yields significant cost reduction while maintaining complete critical infrastructure coverage. For small problems, exact methods dominate in both cost and computation time; for larger problems, the proposed heuristic achieves feasible solutions in under 3 minutes with approximately 14% optimality gap where comparison is possible. From systems perspective, our work demonstrate that optimized planning for human-robot teaming is key to achieve both cost-effective and mission-reliable deployments.", "date": "2026-02-08", "categories": []} {"id": "2604.13283", "title": "Optimizing Earth Observation Satellite Schedules under Unknown Operational Constraints: An Active Constraint Acquisition Approach", "abstract": "Earth Observation (EO) satellite scheduling (deciding which imaging tasks to perform and when) is a well-studied combinatorial optimization problem. Existing methods typically assume that the operational constraint model is fully specified in advance. In practice, however, constraints governing separation between observations, power budgets, and thermal limits are often embedded in engineering artefacts or high-fidelity simulators rather than in explicit mathematical models. We study EO scheduling under \\emph{unknown constraints}: the objective is known, but feasibility must be learned interactively from a binary oracle. Working with a simplified model restricted to pairwise separation and global capacity constraints, we introduce Conservative Constraint Acquisition~(CCA), a domain-specific procedure designed to identify justified constraints efficiently in practice while limiting unnecessary tightening of the learned model. Embedded in the \\textsc{Learn\\&Optimize} framework, CCA supports an interactive search process that alternates optimization under a learned constraint model with targeted oracle queries. On synthetic instances with up to 50~tasks and dense constraint networks, L\\&O improves over a no-knowledge greedy baseline and uses far fewer main oracle queries than a two-phase acquire-then-solve baseline (FAO). For $n\\leq 30$, the average gap drops from 65--68\\% (Priority Greedy) to 17.7--35.8\\% using L\\&O. At $n{=}50$, where the CP-SAT reference is the best feasible solution found in 120~s, L\\&O improves on FAO on average (17.9\\% vs.\\ 20.3\\%) while using 21.3 main queries instead of 100 and about $5\\times$ less execution time.", "date": "2026-04-14", "categories": []} {"id": "2604.13385", "title": "On the Use of Evolutionary Optimization for the Dynamic Chance Constrained Open-Pit Mine Scheduling Problem", "abstract": "Open-pit mine scheduling is a complex real world optimization problem that involves uncertain economic values and dynamically changing resource capacities. Evolutionary algorithms are particularly effective in these scenarios, as they can easily adapt to uncertain and changing environments. However, uncertainty and dynamic changes are often studied in isolation in real-world problems. In this paper, we study a dynamic chance-constrained open-pit mine scheduling problem in which block economic values are stochastic and mining and processing capacities vary over time. We adopt a bi-objective evolutionary formulation that simultaneously maximizes expected discounted profit and minimizes its standard deviation. To address dynamic changes, we propose a diversity-based change response mechanism that repairs a subset of infeasible solutions and introduces additional feasible solutions whenever a change is detected. We evaluate the effectiveness of this mechanism across four multi-objective evolutionary algorithms and compare it with a baseline re-evaluation-based change-response strategy. Experimental results on six mining instances demonstrate that the proposed approach consistently outperforms the baseline methods across different uncertainty levels and change frequencies.", "date": "2026-04-15", "categories": []} {"id": "2604.01455", "title": "Infeasibility Aware Large Language Models for Combinatorial Optimization", "abstract": "Large language models (LLMs) are increasingly explored for NP-hard combinatorial optimization problems, but most existing methods emphasize feasible-instance solution generation and do not explicitly address infeasibility detection. We propose an infeasibility-aware framework that combines certifiable dataset construction, supervised fine-tuning, and LLM-assisted downstream search. For the minor-embedding problem, we introduce a new mathematical programming formulation together with provable zero-phase infeasibility screening, which enables scalable construction of training instances labeled either as feasible with structured certificates or as certifiably infeasible. Using training data generated through this exact optimization pipeline, we show that an 8B-parameter LLM can be fine-tuned to jointly perform solution generation and infeasibility detection. We further utilize LLM outputs as warm starts for downstream local search, providing a practical way to accelerate optimization even when the LLM outputs are imperfect. Experiments show that our fine-tuned model improves overall accuracy by up to 30\\% over GPT-5.2; meanwhile LLM-guided warm starts provide up to $2\\times$ speedup compared with starting from scratch in downstream local search.", "date": "2026-04-01", "categories": []} {"id": "2604.12898", "title": "BEAM: Bi-level Memory-adaptive Algorithmic Evolution for LLM-Powered Heuristic Design", "abstract": "Large Language Model-based Hyper Heuristic (LHH) has recently emerged as an efficient way for automatic heuristic design. However, most existing LHHs just perform well in optimizing a single function within a pre-defined solver. Their single-layer evolution makes them not effective enough to write a competent complete solver. While some variants incorporate hyperparameter tuning or attempt to generate complex code through iterative local modifications, they still lack a high-level algorithmic modeling, leading to limited exploration efficiency. To address this, we reformulate heuristic design as a Bi-level Optimization problem and propose \\textbf{BEAM} (Bi-level Memory-adaptive Algorithmic Evolution). BEAM's exterior layer evolves high-level algorithmic structures with function placeholders through genetic algorithm (GA), while the interior layer realizes these placeholders via Monte Carlo Tree Search (MCTS). We further introduce an Adaptive Memory module to facilitate complex code generation. To support the evaluation for complex code generation, we point out the limitations of starting LHHs from scratch or from code templates and introduce a Knowledge Augmentation (KA) Pipeline. Experimental results on several optimization problems demonstrate that BEAM significantly outperforms existing LHHs, notably reducing the optimality gap by 37.84\\% on aggregate in CVRP hybrid algorithm design. BEAM also designs a heuristic that outperforms SOTA Maximum Independent Set (MIS) solver KaMIS.", "date": "2026-04-14", "categories": []} {"id": "2604.13013", "title": "Bilevel Late Acceptance Hill Climbing for the Electric Capacitated Vehicle Routing Problem", "abstract": "This paper tackles the Electric Capacitated Vehicle Routing Problem (E-CVRP) through a bilevel optimization framework that handles routing and charging decisions separately or jointly depending on the search stage. By analyzing their interaction, we introduce a surrogate objective at the upper level to guide the search and accelerate convergence. A bilevel Late Acceptance Hill Climbing algorithm (b-LAHC) is introduced that operates through three phases: greedy descent, neighborhood exploration, and final solution refinement. b-LAHC operates with fixed parameters, eliminating the need for complex adaptation while remaining lightweight and effective. Extensive experiments on the IEEE WCCI-2020 benchmark show that b-LAHC achieves superior or competitive performance against eight state-of-the-art algorithms. Under a fixed evaluation budget, it attains near-optimal solutions on small-scale instances and sets 9/10 new best-known results on large-scale benchmarks, improving existing records by an average of 1.07%. Moreover, the strong correlation (though not universal) observed between the surrogate objective and the complete cost justifies the use of the surrogate objective while still necessitating a joint solution of both levels, thereby validating the effectiveness of the proposed bilevel framework and highlighting its potential for efficiently solving large-scale routing problems with a hierarchical structure.", "date": "2026-04-14", "categories": []} {"id": "2510.22329", "title": "Graph-Coarsening Approach for the Capacitated Vehicle Routing Problem with Time Windows", "abstract": "The Capacitated Vehicle Routing Problem with Time Windows (CVRPTW) is a fundamental NP-hard optimization problem in logistics. Solving large-scale instances remains computationally challenging for exact solvers. This paper introduces a multilevel graph coarsening and refinement strategy that aggregates customers into meta-nodes based on a spatio-temporal distance metric. The reduced problem is solved using both classical heuristics and quantum annealing hardware, then expanded back into the original space with arrival times recomputed and constraint violations recorded. Comprehensive experiments on Solomon benchmarks demonstrate that our method significantly reduces computation time while preserving solution quality for classical heuristics. For quantum solvers, experiments across all 56 Solomon instances at $N=5$ and $N=10$ customers show that coarsening consistently reduces computation time and, on clustered (C-type) instances, simultaneously reduces vehicle count and route duration with no feasibility loss. Coarsening effectiveness is strongly instance-structure dependent: C-type instances achieve %100 post-coarsening feasibility with measurable quality improvements, while narrow-window random (R-type) instances present structural constraints that limit achievable coarsening depth.", "date": "2025-10-25", "categories": []} {"id": "2604.09556", "title": "Para-B&B: Load-Balanced Deterministic Parallelization of Solving MIP", "abstract": "Mixed-integer programming (MIP) extends linear programming by incorporating both continuous and integer decision variables, making it widely used in production planning, logistics scheduling, and resource allocation. However, MIP remains NP-hard and cannot generally be solved to optimality in polynomial time. Branch-and-bound, a fundamental exact method, faces significant parallelization challenges due to computational heterogeneity and strict determinism requirements in commercial applications. This paper presents the first fully open-source implementation of deterministic parallel branch-and-bound for HiGHS, a high-performance MIP solver. Our approach introduces a novel data-parallel architecture ensuring strict determinism by replicating complete solver state across worker threads and eliminating non-deterministic synchronization primitives. A key innovation is our AI-driven load balancing mechanism employing multi-stage workload prediction models that estimate node computational complexity based on structural characteristics and historical performance data, coupled with dynamic parameter adjustment strategies. The framework executes orchestrated parallel phases including concurrent dive operations, systematic data consolidation, and intelligent node selection. Comprehensive experimental evaluation on 80 MIPLIB 2017 benchmark instances demonstrates effectiveness, achieving a geometric mean speedup of 2.17 using eight threads while maintaining complete deterministic guarantees. Performance gains become increasingly pronounced for higher node counts, with speedup factors reaching 5.12 for computationally intensive instances and thread idle rates averaging 34.7%.", "date": "2026-02-10", "categories": []} {"id": "2604.10073", "title": "Graph-RHO: Critical-path-aware Heterogeneous Graph Network for Long-Horizon Flexible Job-Shop Scheduling", "abstract": "Long-horizon Flexible Job-Shop Scheduling~(FJSP) presents a formidable combinatorial challenge due to complex, interdependent decisions spanning extended time horizons. While learning-based Rolling Horizon Optimization~(RHO) has emerged as a promising paradigm to accelerate solving by identifying and fixing invariant operations, its effectiveness is hindered by the structural complexity of FJSP. Existing methods often fail to capture intricate graph-structured dependencies and ignore the asymmetric costs of prediction errors, in which misclassifying critical-path operations is significantly more detrimental than misclassifying non-critical ones. Furthermore, dynamic shifts in predictive confidence during the rolling process make static pruning thresholds inadequate. To address these limitations, we propose Graph-RHO, a novel critical-path-aware graph-based RHO framework. First, we introduce a topology-aware heterogeneous graph network that encodes subproblems as operation-machine graphs with multi-relational edges, leveraging edge-feature-aware message passing to predict operation stability. Second, we incorporate a critical-path-aware mechanism that injects inductive biases during training to distinguish highly sensitive bottleneck operations from robust ones. Third, we devise an adaptive thresholding strategy that dynamically calibrates decision boundaries based on online uncertainty estimation to align model predictions with the solver's search space. Extensive experiments on standard benchmarks demonstrate that \\mbox{Graph-RHO} establishes a new state of the art in solution quality and computational efficiency. Remarkably, it exhibits exceptional zero-shot generalization, reducing solve time by over 30\\% on large-scale instances (2000 operations) while achieving superior solution quality. Our code is available \\href{https://github.com/IntelliSensing/Graph-RHO}{here}.", "date": "2026-04-11", "categories": []} {"id": "2604.10652", "title": "Enhancing Cross-Problem Vehicle Routing via Federated Learning", "abstract": "Vehicle routing problems (VRPs) constitute a core optimization challenge in modern logistics and supply chain management. The recent neural combinatorial optimization (NCO) has demonstrated superior efficiency over some traditional algorithms. While serving as a primary NCO approach for solving general VRPs, current cross-problem learning paradigms are still subject to performance degradation and generalizability decay, when transferring from simple VRP variants to those involving different and complex constraints. To strengthen the paradigms, this paper offers an innovative \"Multi-problem Pre-train, then Single-problem Fine-tune\" framework with Federated Learning (MPSF-FL). This framework exploits the common knowledge of a federated global model to foster efficient cross-problem knowledge sharing and transfer among local models for single-problem fine-tuning. In this way, local models effectively retain common VRP knowledge from up-to-date global model, while being efficiently adapted to downstream VRPs with heterogeneous complex constraints. Experimental results demonstrate that our framework not only enhances the performance in diverse VRPs, but also improves the generalizability in unseen problems.", "date": "2026-04-12", "categories": []} {"id": "2602.02188", "title": "Reasoning in a Combinatorial and Constrained World: Benchmarking LLMs on Natural-Language Combinatorial Optimization", "abstract": "While large language models (LLMs) have shown strong performance in math and logic reasoning, their ability to handle combinatorial optimization (CO) -- searching high-dimensional solution spaces under hard constraints -- remains underexplored. To bridge the gap, we introduce NLCO, a \\textbf{N}atural \\textbf{L}anguage \\textbf{C}ombinatorial \\textbf{O}ptimization benchmark that evaluates LLMs on end-to-end CO reasoning: given a language-described decision-making scenario, the model must output a discrete solution without writing code or calling external solvers. NLCO covers 43 CO problems and is organized using a four-layer taxonomy of variable types, constraint families, global patterns, and objective classes, enabling fine-grained evaluation. We provide solver-annotated solutions and comprehensively evaluate LLMs by feasibility, solution optimality, and reasoning efficiency. Experiments across a wide range of modern LLMs show that high-performing models achieve strong feasibility and solution quality on small instances, but both degrade as instance size grows, even if more tokens are used for reasoning. We also observe systematic effects across the taxonomy: set-based tasks are relatively easy, whereas graph-structured problems and bottleneck objectives lead to more frequent failures.", "date": "2026-02-02", "categories": []} {"id": "2604.07872", "title": "PyVRP$^+$: LLM-Driven Metacognitive Heuristic Evolution for Hybrid Genetic Search in Vehicle Routing Problems", "abstract": "Designing high-performing metaheuristics for NP-hard combinatorial optimization problems, such as the Vehicle Routing Problem (VRP), remains a significant challenge, often requiring extensive domain expertise and manual tuning. Recent advances have demonstrated the potential of large language models (LLMs) to automate this process through evolutionary search. However, existing methods are largely reactive, relying on immediate performance feedback to guide what are essentially black-box code mutations. Our work departs from this paradigm by introducing Metacognitive Evolutionary Programming (MEP), a framework that elevates the LLM to a strategic discovery agent. Instead of merely reacting to performance scores, MEP compels the LLM to engage in a structured Reason-Act-Reflect cycle, forcing it to explicitly diagnose failures, formulate design hypotheses, and implement solutions grounded in pre-supplied domain knowledge. By applying MEP to evolve core components of the state-of-the-art Hybrid Genetic Search (HGS) algorithm, we discover novel heuristics that significantly outperform the original baseline. By steering the LLM to reason strategically about the exploration-exploitation trade-off, our approach discovers more effective and efficient heuristics applicable across a wide spectrum of VRP variants. Our results show that MEP discovers heuristics that yield significant performance gains over the original HGS baseline, improving solution quality by up to 2.70\\% and reducing runtime by over 45\\% on challenging VRP variants.", "date": "2026-04-09", "categories": []} {"id": "2403.06568", "title": "Better Understandings and Configurations in MaxSAT Local Search Solvers via Anytime Performance Analysis", "abstract": "Though numerous solvers have been proposed for the MaxSAT problem, and the benchmark environment such as MaxSAT Evaluations provides a platform for the comparison of the state-of-the-art solvers, existing assessments were usually evaluated based on the quality, e.g., fitness, of the best-found solutions obtained within a given running time budget. However, concerning solely the final obtained solutions regarding specific time budgets may restrict us from comprehending the behavior of the solvers along the convergence process. This paper demonstrates that Empirical Cumulative Distribution Functions can be used to compare MaxSAT stochastic local search solvers' anytime performance across multiple problem instances and various time budgets. The assessment reveals distinctions in solvers' performance and displays that the (dis)advantages of solvers adjust along different running times. This work also exhibits that the quantitative and high variance assessment of anytime performance can guide machines, i.e., automatic configurators, to search for better parameter settings. Our experimental results show that the hyperparameter optimization tool, i.e., SMAC, can achieve better parameter settings of solvers when using the anytime performance as the cost function, compared to using the metrics based on the fitness of the best-found solutions.", "date": "2024-03-11", "categories": []} {"id": "2604.06724", "title": "The Traveling Thief Problem with Time Windows: Benchmarks and Heuristics", "abstract": "While traditional optimization problems were often studied in isolation, many real-world problems today require interdependence among multiple optimization components. The traveling thief problem (TTP) is a multi-component problem that has been widely studied in the literature. In this paper, we introduce and investigate the TTP with time window constraints which provides a TTP variant highly relevant to real-world situations where good can only be collected at given time intervals. We examine adaptions of existing approaches for TTP and the Traveling Salesperson Problem (TSP) with time windows to this new problem and evaluate their performance. Furthermore, we provide a new heuristic approach for the TTP with time windows. To evaluate algorithms for TTP with time windows, we introduce new TTP benchmark instances with time windows based on TTP instances existing in the literature. Our experimental investigations evaluate the different approaches and show that the newly designed algorithm outperforms the other approaches on a wide range of benchmark instances.", "date": "2026-04-08", "categories": []} {"id": "2406.14514", "title": "Solving a Stackelberg Game on Transportation Networks in a Dynamic Crime Scenario: A Mixed Approach on Multi-Layer Networks", "abstract": "Interdicting a criminal with limited police resources is a challenging task as the criminal changes location over time. The size of the large transportation network further adds to the difficulty of this scenario. To tackle this issue, we consider the concept of a layered graph. At each time stamp, we create a copy of the entire transportation network to track the possible movements of both players, the attacker and the defenders. We consider a Stackelberg game in a dynamic crime scenario where the attacker changes location over time while the defenders attempt to interdict the attacker on his escape route. Given a set of defender strategies, the optimal attacker strategy is determined by applying Dijkstra's algorithm on the layered networks. Here, the attacker aims to minimize while the defenders aim to maximize the probability of interdiction. We develop an approximation algorithm on the layered networks to find near-optimal strategy for defenders. The efficacy of the developed approach is compared with the adopted MILP approach. We compare the results in terms of computational time and solution quality. The quality of the results demonstrates the need for the developed approach, as it effectively solves the complex problem within a short amount of time.", "date": "2024-06-20", "categories": []} {"id": "2511.10809", "title": "Near-optimal Linear Predictive Clustering in Non-separable Spaces via MIP and QPBO Reductions", "abstract": "Linear Predictive Clustering (LPC) partitions samples based on shared linear relationships between feature and target variables, with numerous applications including marketing, medicine, and education. Greedy optimization methods, commonly used for LPC, alternate between clustering and linear regression but lack global optimality. While effective for separable clusters, they struggle in non-separable settings where clusters overlap in feature space. In an alternative constrained optimization paradigm, Bertsimas and Shioda (2007) formulated LPC as a Mixed-Integer Program (MIP), ensuring global optimality regardless of separability but suffering from poor scalability. This work builds on the constrained optimization paradigm to introduce two novel approaches that improve the efficiency of global optimization for LPC. By leveraging key theoretical properties of separability, we derive near-optimal approximations with provable error bounds, significantly reducing the MIP formulation's complexity and improving scalability. Additionally, we can further approximate LPC as a Quadratic Pseudo-Boolean Optimization (QPBO) problem, achieving substantial computational improvements in some settings. Comparative analyses on synthetic and real-world datasets demonstrate that our methods consistently achieve near-optimal solutions with substantially lower regression errors than greedy optimization while exhibiting superior scalability over existing MIP formulations.", "date": "2025-11-13", "categories": []} {"id": "2604.04941", "title": "Algebraic Structure Discovery for Real World Combinatorial Optimisation Problems: A General Framework from Abstract Algebra to Quotient Space Learning", "abstract": "Many combinatorial optimisation problems hide algebraic structures that, once exposed, shrink the search space and improve the chance of finding the global optimal solution. We present a general framework that (i) identifies algebraic structure, (ii) formalises operations, (iii) constructs quotient spaces that collapse redundant representations, and (iv) optimises directly over these reduced spaces. Across a broad family of rule-combination tasks (e.g., patient subgroup discovery and rule-based molecular screening), conjunctive rules form a monoid. Via a characteristic-vector encoding, we prove an isomorphism to the Boolean hypercube $\\{0,1\\}^n$ with bitwise OR, so logical AND in rules becomes bitwise OR in the encoding. This yields a principled quotient-space formulation that groups functionally equivalent rules and guides structure-aware search. On real clinical data and synthetic benchmarks, quotient-space-aware genetic algorithms recover the global optimum in 48% to 77% of runs versus 35% to 37% for standard approaches, while maintaining diversity across equivalence classes. These results show that exposing and exploiting algebraic structure offers a simple, general route to more efficient combinatorial optimisation.", "date": "2026-03-10", "categories": []} {"id": "2604.05195", "title": "Vehicle-as-Prompt: A Unified Deep Reinforcement Learning Framework for Heterogeneous Fleet Vehicle Routing Problem", "abstract": "Unlike traditional homogeneous routing problems, the Heterogeneous Fleet Vehicle Routing Problem (HFVRP) involves heterogeneous fixed costs, variable travel costs, and capacity constraints, rendering solution quality highly sensitive to vehicle selection. Furthermore, real-world logistics applications often impose additional complex constraints, markedly increasing computational complexity. However, most existing Deep Reinforcement Learning (DRL)-based methods are restricted to homogeneous scenarios, leading to suboptimal performance when applied to HFVRP and its complex variants. To bridge this gap, we investigate HFVRP under complex constraints and develop a unified DRL framework capable of solving the problem across various variant settings. We introduce the Vehicle-as-Prompt (VaP) mechanism, which formulates the problem as a single-stage autoregressive decision process. Building on this, we propose VaP-CSMV, a framework featuring a cross-semantic encoder and a multi-view decoder that effectively addresses various problem variants and captures the complex mapping relationships between vehicle heterogeneity and customer node attributes. Extensive experimental results demonstrate that VaP-CSMV significantly outperforms existing state-of-the-art DRL-based neural solvers and achieves competitive solution quality compared to traditional heuristic solvers, while reducing inference time to mere seconds. Furthermore, the framework exhibits strong zero-shot generalization capabilities on large-scale and previously unseen problem variants, while ablation studies validate the vital contribution of each component.", "date": "2026-04-06", "categories": []} {"id": "2604.05416", "title": "Multi-Agent Pathfinding with Non-Unit Integer Edge Costs via Enhanced Conflict-Based Search and Graph Discretization", "abstract": "Multi-Agent Pathfinding (MAPF) plays a critical role in various domains. Traditional MAPF methods typically assume unit edge costs and single-timestep actions, which limit their applicability to real-world scenarios. MAPFR extends MAPF to handle non-unit costs with real-valued edge costs and continuous-time actions, but its geometric collision model leads to an unbounded state space that compromises solver efficiency. In this paper, we propose MAPFZ, a novel MAPF variant on graphs with non-unit integer costs that preserves a finite state space while offering improved realism over classical MAPF. To solve MAPFZ efficiently, we develop CBS-NIC, an enhanced Conflict-Based Search framework incorporating time-interval-based conflict detection and an improved Safe Interval Path Planning (SIPP) algorithm. Additionally, we propose Bayesian Optimization for Graph Design (BOGD), a discretization method for non-unit edge costs that balances efficiency and accuracy with a sub-linear regret bound. Extensive experiments demonstrate that our approach outperforms state-of-the-art methods in runtime and success rate across diverse benchmark scenarios.", "date": "2026-04-07", "categories": []} {"id": "2604.05530", "title": "Inventory of the 12 007 Low-Dimensional Pseudo-Boolean Landscapes Invariant to Rank, Translation, and Rotation", "abstract": "Many randomized optimization algorithms are rank-invariant, relying solely on the relative ordering of solutions rather than absolute fitness values. We introduce a stronger notion of rank landscape invariance: two problems are equivalent if their ranking, but also their neighborhood structure and symmetries (translation and rotation), induce identical landscapes. This motivates the study of rank landscapes rather than individual functions. While prior work analyzed the rankings of injective function classes in isolation, we provide an exhaustive inventory of the invariant landscape classes for pseudo-Boolean functions of dimensions 1, 2, and 3, including non-injective cases. Our analysis reveals 12,007 classes in total, a significant reduction compared to rank-invariance alone. We find that non-injective functions yield far more invariant landscape classes than injective ones. In addition, complex combinations of topological landscape properties and algorithm behaviors emerge, particularly regarding deceptiveness, neutrality, and the performance of hill-climbing strategies. The inventory serves as a resource for pedagogical purposes and benchmark design, offering a foundation for constructing larger problems with controlled hardness and advancing our understanding of landscape difficulty and algorithm performance.", "date": "2026-04-07", "categories": []} {"id": "2411.14411", "title": "Multi-Agent Environments for Vehicle Routing Problems", "abstract": "Research on Reinforcement Learning (RL) approaches for discrete optimization problems has increased considerably, extending RL to areas classically dominated by Operations Research (OR). Vehicle routing problems are a good example of discrete optimization problems with high practical relevance, for which RL techniques have achieved notable success. Despite these advances, open-source development frameworks remain scarce, hindering both algorithm testing and objective comparison of results. This situation ultimately slows down progress in the field and limits the exchange of ideas between the RL and OR communities. Here, we propose MAEnvs4VRP library, a unified framework for multi-agent vehicle routing environments that supports classical, dynamic, stochastic, and multi-task problem variants within a single modular design. The library, built on PyTorch, provides a flexible and modular architecture design that facilitates customization and the incorporation of new routing problems. It follows the Agent Environment Cycle (\"AEC\") games model and features an intuitive API, enabling rapid adoption and seamless integration into existing reinforcement learning frameworks. The project source code can be found at https://github.com/ricgama/maenvs4vrp.", "date": "2024-11-21", "categories": []} {"id": "2506.09749", "title": "Large Language Models for Combinatorial Optimization of Design Structure Matrix", "abstract": "In complex engineering systems, the dependencies among components or development activities are often modeled and analyzed using Design Structure Matrix (DSM). Reorganizing elements within a DSM to minimize feedback loops and enhance modularity or process efficiency constitutes a challenging combinatorial optimization (CO) problem in engineering design and operations. As problem sizes increase and dependency networks become more intricate, traditional optimization methods that rely solely on mathematical heuristics often fail to capture the contextual nuances and struggle to deliver effective solutions. In this study, we explore the potential of Large Language Models (LLMs) to address such CO problems by leveraging their capabilities for advanced reasoning and contextual understanding. We propose a novel LLM-based framework that integrates network topology with contextual domain knowledge for iterative optimization of DSM sequencing-a common CO problem. Experiments on various DSM cases demonstrate that our method consistently achieves faster convergence and superior solution quality compared to both stochastic and deterministic baselines. Notably, incorporating contextual domain knowledge significantly enhances optimization performance regardless of the chosen LLM backbone. These findings highlight the potential of LLMs to solve complex engineering CO problems by combining semantic and mathematical reasoning. This approach paves the way towards a new paradigm in LLM-based engineering design optimization.", "date": "2025-06-11", "categories": []} {"id": "2604.03234", "title": "Structural Segmentation of the Minimum Set Cover Problem: Exploiting Universe Decomposability for Metaheuristic Optimization", "abstract": "The Minimum Set Cover Problem (MSCP) is a classical NP-hard combinatorial optimization problem with numerous applications in science and engineering. Although a wide range of exact, approximate, and metaheuristic approaches have been proposed, most methods implicitly treat MSCP instances as monolithic, overlooking potential intrinsic structural properties of the universe. In this work, we investigate the concept of \\emph{universe segmentability} in the MSCP and analyze how intrinsic structural decomposition (universe segmentability) can be exploited to enhance heuristic optimization. We propose an efficient preprocessing strategy based on disjoint-set union (union--find) to detect connected components induced by element co-occurrence within subsets, enabling the decomposition of the original instance into independent subproblems. Each subproblem is solved using the GRASP metaheuristic, and partial solutions are combined without compromising feasibility. Extensive experiments on standard benchmark instances and large-scale synthetic datasets show that exploiting natural universe segmentation consistently improves solution quality and scalability, particularly for large and structurally decomposable instances. These gains are supported by a succinct bit-level set representation that enables efficient set operations, making the proposed approach computationally practical at scale.", "date": "2026-01-29", "categories": []} {"id": "2505.17938", "title": "LMask: Learn to Solve Constrained Routing Problems with Lazy Masking", "abstract": "Routing problems are canonical combinatorial optimization tasks with wide-ranging applications in logistics, transportation, and supply chain management. However, solving these problems becomes significantly more challenging when complex constraints are involved. In this paper, we propose LMask, a novel learning framework that utilizes dynamic masking to generate high-quality feasible solutions for constrained routing problems. LMask introduces the LazyMask decoding method, which lazily refines feasibility masks with the backtracking mechanism. In addition, it employs the refinement intensity embedding to encode the search trace into the model, mitigating representation ambiguities induced by backtracking. To further reduce sampling cost, LMask sets a backtracking budget during decoding, while constraint violations are penalized in the loss function during training to counteract infeasibility caused by this budget. We provide theoretical guarantees for the validity and probabilistic optimality of our approach. Extensive experiments on the traveling salesman problem with time windows (TSPTW) and TSP with draft limits (TSPDL) demonstrate that LMask achieves state-of-the-art feasibility rates and solution quality, outperforming existing neural methods.", "date": "2025-05-23", "categories": []} {"id": "2509.12643", "title": "Learn to Relax with Large Language Models: Solving Constraint Optimization Problems via Bidirectional Coevolution", "abstract": "Large Language Model (LLM)-based optimization has recently shown promise for autonomous problem solving, yet most approaches still cast LLMs as passive constraint checkers rather than proactive strategy designers, limiting their effectiveness on complex Constraint Optimization Problems (COPs). To address this, we present AutoCO, an end-to-end Automated Constraint Optimization method that tightly couples operations-research principles of constraint relaxation with LLM reasoning. A core innovation is a unified triple-representation that binds relaxation strategies, algorithmic principles, and executable codes. This design enables the LLM to synthesize, justify, and instantiate relaxation strategies that are both principled and executable. To navigate fragmented solution spaces, AutoCO employs a bidirectional global-local coevolution mechanism, synergistically coupling Monte Carlo Tree Search (MCTS) for global relaxation-trajectory exploration with Evolutionary Algorithms (EAs) for local solution intensification. This continuous exchange of priors and feedback explicitly balances diversification and intensification, thus preventing premature convergence. Extensive experiments on three challenging COP benchmarks validate AutoCO's consistent effectiveness and superior performance, especially in hard regimes where current methods degrade. Results highlight AutoCO as a principled and effective path toward proactive, verifiable LLM-driven optimization.", "date": "2025-09-16", "categories": []} {"id": "2604.01732", "title": "Solving the Two-dimensional single stock size Cutting Stock Problem with SAT and MaxSAT", "abstract": "Cutting rectangular items from stock sheets to satisfy demands while minimizing waste is a central manufacturing task. The Two-Dimensional Single Stock Size Cutting Stock Problem (2D-CSSP) generalizes bin packing by requiring multiple copies of each item type, which causes a strong combinatorial blow-up. We present a SAT-based framework where item types are expanded by demand, each copy has a sheet-assignment variable and non-overlap constraints are activated only for copies assigned to the same sheet. We also introduce an infeasible-orientation elimination rule that fixes rotation variables when only one orientation can fit the sheet. For minimizing the number of sheets, we compare three approaches: non-incremental SAT with binary search, incremental SAT with clause reuse across iterations and weighted partial MaxSAT. On the Cui--Zhao benchmark suite, our best SAT configurations certify two to three times more instances as provably optimal and achieve lower optimality gaps than OR-Tools, CPLEX and Gurobi. The relative ranking among SAT approaches depends on rotation: incremental SAT is strongest without rotation, while non-incremental SAT is more effective when rotation increases formula size.", "date": "2026-04-02", "categories": []} {"id": "2411.04685", "title": "Solving Generalized Grouping Problems in Cellular Manufacturing Systems Using a Network Flow Model", "abstract": "This paper focuses on the generalized grouping problem in the context of cellular manufacturing systems (CMS), where parts may have more than one process route. A process route lists the machines corresponding to each part of the operation. Inspired by the extensive and widespread use of network flow algorithms, this research formulates the process route family formation for generalized grouping as a unit capacity minimum cost network flow model. The objective is to minimize dissimilarity (based on the machines required) among the process routes within a family. The proposed model optimally solves the process route family formation problem without pre-specifying the number of part families to be formed. The process route of family formation is the first stage in a hierarchical procedure. For the second stage (machine cell formation), two procedures, a quadratic assignment programming (QAP) formulation, and a heuristic procedure, are proposed. The QAP simultaneously assigns process route families and machines to a pre-specified number of cells in such a way that total machine utilization is maximized. The heuristic procedure for machine cell formation is hierarchical in nature. Computational results for some test problems show that the QAP and the heuristic procedure yield the same results.", "date": "2024-11-07", "categories": []} {"id": "2511.09219", "title": "Planning in Branch-and-Bound: Model-Based Reinforcement Learning for Exact Combinatorial Optimization", "abstract": "Mixed-Integer Linear Programming (MILP) lies at the core of many real-world combinatorial optimization (CO) problems, traditionally solved by branch-and-bound (B&B). A key driver influencing B&B solvers efficiency is the variable selection heuristic that guides branching decisions. Looking to move beyond static, hand-crafted heuristics, recent work has explored adapting traditional reinforcement learning (RL) algorithms to the B&B setting, aiming to learn branching strategies tailored to specific MILP distributions. In parallel, RL agents have achieved remarkable success in board games, a very specific type of combinatorial problems, by leveraging environment simulators to plan via Monte Carlo Tree Search (MCTS). Building on these developments, we introduce Plan-and-Branch-and-Bound (PlanB&B), a model-based reinforcement learning (MBRL) agent that leverages a learned internal model of the B&B dynamics to discover improved branching strategies. Computational experiments empirically validate our approach, with our MBRL branching agent outperforming previous state-of-the-art RL methods across four standard MILP benchmarks.", "date": "2025-11-12", "categories": []} {"id": "2603.28943", "title": "Differentiable Initialization-Accelerated CPU-GPU Hybrid Combinatorial Scheduling", "abstract": "This paper presents a hybrid CPU-GPU framework for solving combinatorial scheduling problems formulated as Integer Linear Programming (ILP). While scheduling underpins many optimization tasks in computing systems, solving these problems optimally at scale remains a long-standing challenge due to their NP-hard nature. We introduce a novel approach that combines differentiable optimization with classical ILP solving. Specifically, we utilize differentiable presolving to rapidly generate high-quality partial solutions, which serve as warm-starts for commercial ILP solvers (CPLEX, Gurobi) and rising open-source solver HiGHS. This method enables significantly improved early pruning compared to state-of-the-art standalone solvers. Empirical results across industry-scale benchmarks demonstrate up to a $10\\times$ performance gain over baselines, narrowing the optimality gap to $<0.1\\%$. This work represents the first demonstration of utilizing differentiable optimization to initialize exact ILP solvers for combinatorial scheduling, opening new opportunities to integrate machine learning infrastructure with classical exact optimization methods across broader domains.", "date": "2026-03-30", "categories": []} {"id": "2603.29543", "title": "Reducing Complexity for Quantum Approaches in Train Load Optimization", "abstract": "Efficiently planning container loads onto trains is a computationally challenging combinatorial optimization problem, central to logistics and supply chain management. A primary source of this complexity arises from the need to model and reduce rehandle operations-unproductive crane moves required to access blocked containers. Conventional mathematical formulations address this by introducing explicit binary variables and a web of logical constraints for each potential rehandle, resulting in large-scale models that are difficult to solve. This paper presents a fundamental departure from this paradigm. We introduce an innovative and compact mathematical formulation for the Train Load Optimization (TLO) problem where the rehandle cost is calculated implicitly within the objective function. This novel approach helps prevent the need for dedicated rehandle variables and their associated constraints, leading to a dramatic reduction in model size. We provide a formal comparison against a conventional model to analytically demonstrate the significant reduction in the number of variables and constraints. The efficacy of our compact formulation is assessed through a simulated annealing metaheuristic, which finds high-quality loading plans for various problem instances. The results confirm that our model is not only more parsimonious but also practically effective, offering a scalable and powerful tool for modern rail logistics.", "date": "2026-03-31", "categories": []} {"id": "2603.29643", "title": "Optimizing Donor Outreach for Blood Collection Sessions: A Scalable Decision Support Framework", "abstract": "Blood donation centers face challenges in matching supply with demand while managing donor availability. Although targeted outreach is important, it can cause donor fatigue via over-solicitation. Effective recruitment requires targeting the right donors at the right time, balancing constraints with donor convenience and eligibility. Despite extensive work on blood supply chain optimization and growing interest in algorithmic donor recruitment, the operational problem of assigning donors to sessions across a multi-site network, taking into account eligibility, capacity, blood-type demand targets, geographic convenience, and donor safety, remains unaddressed.\n We address this gap with an optimization framework for donor invitation scheduling incorporating donor eligibility, travel convenience, blood-type demand targets, and penalties. We evaluate two strategies: (i) a binary integer linear programming (BILP) formulation and (ii) an efficient greedy heuristic. Evaluation uses the registry from Instituto Portugu\\^es do Sangue e da Transplanta\\c{c}\\~ao (IPST) for invite planning in the Lisbon operational region using 4-month windows. A prospective pipeline integrates organic attendance forecasting, quantile-based demand targets, and residual capacity estimation for forward-looking invitation plans. Results reveal its key role in closing the supply-demand gap in the Lisbon operational region. A controlled comparison shows that the greedy heuristic achieves results comparable to the BILP, with 188x less peak memory and 115x faster runtime; trade-offs include 3.9 pp lower demand fulfillment (86.1% vs. 90.0%), larger donor-session distance, higher adverse-reaction donor exposure, and greater invitation burden per non-high-frequency donor, reflecting local versus global optimization. Experiments assess how constraint-aware scheduling can close gaps by mobilizing eligible inactive/lapsing donors.", "date": "2026-03-31", "categories": []} {"id": "2603.29865", "title": "Wildfire Suppression: Complexity, Models, and Instances", "abstract": "Wildfires cause major losses worldwide, and the frequency of fire-weather conditions is likely to increase in many regions. We study the allocation of suppression resources over time on a graph-based representation of a landscape to slow down fire propagation. Our contributions are theoretical and methodological. First, we prove that this problem and related variants in the literature are NP-complete, including cases without resource-timing constraints. Second, we propose a new mixed-integer programming (MIP) formulation that obtains state-of-the-art results, showing that MIP is a competitive approach contrary to earlier findings. Third, showing that existing benchmarks lack realism and difficulty, we introduce a physics-grounded instance generator based on Rothermel's surface fire spread model. We use these diverse instances to benchmark the literature, identifying the specific conditions where each algorithm succeeds or fails.", "date": "2026-01-17", "categories": []} {"id": "2510.09328", "title": "Randomized HyperSteiner: A Stochastic Delaunay Triangulation Heuristic for the Hyperbolic Steiner Minimal Tree", "abstract": "We study the problem of constructing Steiner Minimal Trees (SMTs) in hyperbolic space. Exact SMT computation is NP-hard, and existing hyperbolic heuristics such as HyperSteiner are deterministic and often get trapped in locally suboptimal configurations. We introduce Randomized HyperSteiner (RHS), a stochastic Delaunay triangulation heuristic that incorporates randomness into the expansion process and refines candidate trees via Riemannian gradient descent optimization. Experiments on synthetic data sets and a real-world single-cell transcriptomic data show that RHS outperforms Minimum Spanning Tree (MST), Neighbour Joining, and vanilla HyperSteiner (HS). In near-boundary configurations, RHS can achieve a 32% reduction in total length over HS, demonstrating its effectiveness and robustness in diverse data regimes.", "date": "2025-10-10", "categories": []} {"id": "2603.27169", "title": "Aligning LLMs with Graph Neural Solvers for Combinatorial Optimization", "abstract": "Recent research has demonstrated the effectiveness of large language models (LLMs) in solving combinatorial optimization problems (COPs) by representing tasks and instances in natural language. However, purely language-based approaches struggle to accurately capture complex relational structures inherent in many COPs, rendering them less effective at addressing medium-sized or larger instances. To address these limitations, we propose AlignOPT, a novel approach that aligns LLMs with graph neural solvers to learn a more generalizable neural COP heuristic. Specifically, AlignOPT leverages the semantic understanding capabilities of LLMs to encode textual descriptions of COPs and their instances, while concurrently exploiting graph neural solvers to explicitly model the underlying graph structures of COP instances. Our approach facilitates a robust integration and alignment between linguistic semantics and structural representations, enabling more accurate and scalable COP solutions. Experimental results demonstrate that AlignOPT achieves state-of-the-art results across diverse COPs, underscoring its effectiveness in aligning semantic and structural representations. In particular, AlignOPT demonstrates strong generalization, effectively extending to previously unseen COP instances.", "date": "2026-03-28", "categories": []} {"id": "2603.27628", "title": "DSevolve: Enabling Real-Time Adaptive Scheduling on Dynamic Shop Floor with LLM-Evolved Heuristic Portfolios", "abstract": "In dynamic manufacturing environments, disruptions such as machine breakdowns and new order arrivals continuously shift the optimal dispatching strategy, making adaptive rule selection essential. Existing LLM-powered Automatic Heuristic Design (AHD) frameworks evolve toward a single elite rule that cannot meet this adaptability demand. To address this, we present DSevolve, an industrial scheduling framework that evolves a quality-diverse portfolio of dispatching rules offline and adaptively deploys them online with second-level response time. Multi-persona seeding and topology-aware evolutionary operators produce a behaviorally diverse rule archive indexed by a MAP-Elites feature space. Upon each disruption event, a probe-based fingerprinting mechanism characterizes the current shop floor state, retrieves high-quality candidate rules from an offline knowledge base, and selects the best one via rapid look-ahead simulation. Evaluated on 500 dynamic flexible job shop instances derived from real industrial data, DSevolve outperforms state-of-the-art AHD frameworks, classical dispatching rules, genetic programming, and deep reinforcement learning, offering a practical and deployable solution for intelligent shop floor scheduling.", "date": "2026-03-29", "categories": []} {"id": "2603.27803", "title": "Distributed Online Submodular Maximization under Communication Delays: A Simultaneous Decision-Making Approach", "abstract": "We provide a distributed online algorithm for multi-agent submodular maximization under communication delays. We are motivated by the future distributed information-gathering tasks in unknown and dynamic environments, where utility functions naturally exhibit the diminishing-returns property, i.e., submodularity. Existing approaches for online submodular maximization either rely on sequential multi-hop communication, resulting in prohibitive delays and restrictive connectivity assumptions, or restrict each agent's coordination to its one-hop neighborhood only, thereby limiting the coordination performance. To address the issue, we provide the Distributed Online Greedy (DOG) algorithm, which integrates tools from adversarial bandit learning with delayed feedback to enable simultaneous decision-making across arbitrary network topologies. We provide the approximation performance of DOG against an optimal solution, capturing the suboptimality cost due to decentralization as a function of the network structure. Our analyses further reveal a trade-off between coordination performance and convergence time, determined by the magnitude of communication delays. By this trade-off, DOG spans the spectrum between the state-of-the-art fully centralized online coordination approach [1] and fully decentralized one-hop coordination approach [2].", "date": "2026-03-29", "categories": []} {"id": "2603.18866", "title": "Conflict-Based Search for Multi Agent Path Finding with Asynchronous Actions", "abstract": "Multi-Agent Path Finding (MAPF) seeks collision-free paths for multiple agents from their respective start locations to their respective goal locations while minimizing path costs. Most existing MAPF algorithms rely on a common assumption of synchronized actions, where the actions of all agents start at the same time and always take a time unit, which may limit the use of MAPF planners in practice. To get rid of this assumption, Continuous-time Conflict-Based Search (CCBS) is a popular approach that can find optimal solutions for MAPF with asynchronous actions (MAPF-AA). However, CCBS has recently been identified to be incomplete due to an uncountably infinite state space created by continuous wait durations. This paper proposes a new method, Conflict-Based Search with Asynchronous Actions (CBS-AA), which bypasses this theoretical issue and can solve MAPF-AA with completeness and solution optimality guarantees. Based on CBS-AA, we also develop conflict resolution techniques to improve the scalability of CBS-AA further. Our test results show that our method can reduce the number of branches by up to 90%.", "date": "2026-03-19", "categories": []} {"id": "2603.25241", "title": "Offline Decision Transformers for Neural Combinatorial Optimization: Surpassing Heuristics on the Traveling Salesman Problem", "abstract": "Combinatorial optimization problems like the Traveling Salesman Problem are critical in industry yet NP-hard. Neural Combinatorial Optimization has shown promise, but its reliance on online reinforcement learning (RL) hampers deployment and underutilizes decades of algorithmic knowledge. We address these limitations by applying the offline RL framework, Decision Transformer, to learn superior strategies directly from datasets of heuristic solutions; it aims to not only to imitate but to synthesize and outperform them. Concretely, we (i) integrate a Pointer Network to handle the instance-dependent, variable action space of node selection, and (ii) employ expectile regression for optimistic conditioning of Return-to-Go, which is crucial for instances with widely varying optimal values. Experiments show that our method consistently produces higher-quality tours than the four classical heuristics it is trained on, demonstrating the potential of offline RL to unlock and exceed the performance embedded in existing domain knowledge.", "date": "2026-03-26", "categories": []} {"id": "2603.23873", "title": "The DeepXube Software Package for Solving Pathfinding Problems with Learned Heuristic Functions and Search", "abstract": "DeepXube is a free and open-source Python package and command-line tool that seeks to automate the solution of pathfinding problems by using machine learning to learn heuristic functions that guide heuristic search algorithms tailored to deep neural networks (DNNs). DeepXube is comprised of the latest advances in deep reinforcement learning, heuristic search, and formal logic for solving pathfinding problems. This includes limited-horizon Bellman-based learning, hindsight experience replay, batched heuristic search, and specifying goals with answer-set programming. A robust multiple-inheritance structure simplifies the definition of pathfinding domains and the generation of training data. Training heuristic functions is made efficient through the automatic parallelization of the generation of training data across central processing units (CPUs) and reinforcement learning updates across graphics processing units (GPUs). Pathfinding algorithms that take advantage of the parallelism of GPUs and DNN architectures, such as batch weighted A* and Q* search and beam search are easily employed to solve pathfinding problems through command-line arguments. Finally, several convenient features for visualization, code profiling, and progress monitoring during training and solving are available. The GitHub repository is publicly available at https://github.com/forestagostinelli/deepxube.", "date": "2026-03-25", "categories": []} {"id": "2505.20881", "title": "Generalizable Heuristic Generation Through LLMs with Meta-Optimization", "abstract": "Heuristic design with large language models (LLMs) has emerged as a promising approach for tackling combinatorial optimization problems (COPs). However, existing approaches often rely on manually predefined evolutionary computation (EC) heuristic-optimizers and single-task training schemes, which may constrain the exploration of diverse heuristic algorithms and hinder the generalization of the resulting heuristics. To address these issues, we propose Meta-Optimization of Heuristics (MoH), a novel framework that operates at the optimizer level, discovering effective heuristic-optimizers through the principle of meta-learning. Specifically, MoH leverages LLMs to iteratively refine a meta-optimizer that autonomously constructs diverse heuristic-optimizers through (self-)invocation, thereby eliminating the reliance on a predefined EC heuristic-optimizer. These constructed heuristic-optimizers subsequently evolve heuristics for downstream tasks, enabling broader heuristic exploration. Moreover, MoH employs a multi-task training scheme to promote its generalization capability. Experiments on classic COPs demonstrate that MoH constructs an effective and interpretable meta-optimizer, achieving state-of-the-art performance across various downstream tasks, particularly in cross-size settings. Our code is available at: https://github.com/yiding-s/MoH.", "date": "2025-05-27", "categories": []} {"id": "2603.23405", "title": "Planning over MAPF Agent Dependencies via Multi-Dependency PIBT", "abstract": "Modern Multi-Agent Path Finding (MAPF) algorithms must plan for hundreds to thousands of agents in congested environments within a second, requiring highly efficient algorithms. Priority Inheritance with Backtracking (PIBT) is a popular algorithm capable of effectively planning in such situations. However, PIBT is constrained by its rule-based planning procedure and lacks generality because it restricts its search to paths that conflict with at most one other agent. This limitation also applies to Enhanced PIBT (EPIBT), a recent extension of PIBT. In this paper, we describe a new perspective on solving MAPF by planning over agent dependencies. Taking inspiration from PIBT's priority inheritance logic, we define the concept of agent dependencies and propose Multi-Dependency PIBT (MD-PIBT) that searches over agent dependencies. MD-PIBT is a general framework where specific parameterizations can reproduce PIBT and EPIBT. At the same time, alternative configurations yield novel planning strategies that are not expressible by PIBT or EPIBT. Our experiments demonstrate that MD-PIBT effectively plans for as many as 10,000 homogeneous agents under various kinodynamic constraints, including pebble motion, rotation motion, and differential drive robots with speed and acceleration limits. We perform thorough evaluations on different variants of MAPF and find that MD-PIBT is particularly effective in MAPF with large agents.", "date": "2026-03-24", "categories": []} {"id": "2603.21492", "title": "Multinoulli Extension: A Lossless Continuous Relaxation for Partition-Constrained Subset Selection", "abstract": "Identifying the most representative subset for a close-to-submodular objective while satisfying the predefined partition constraint is a fundamental task with numerous applications in machine learning. However, the existing distorted local-search methods are often hindered by their prohibitive query complexities and the rigid requirement for prior knowledge of difficult-to-obtain structural parameters. To overcome these limitations, we introduce a novel algorithm titled Multinoulli-SCG, which not only is parameter-free, but also can achieve the same approximation guarantees as the distorted local-search methods with significantly fewer function evaluations. More specifically, when the objective function is monotone $\\alpha$-weakly DR-submodular or $(\\gamma,\\beta)$-weakly submodular, our Multinoulli-SCG algorithm can attain a value of $(1-e^{-\\alpha})\\text{OPT}-\\epsilon$ or $(\\frac{\\gamma^{2}(1-e^{-(\\beta(1-\\gamma)+\\gamma^2)})}{\\beta(1-\\gamma)+\\gamma^2})\\text{OPT}-\\epsilon$ with only $O(1/\\epsilon^{2})$ function evaluations, where OPT denotes the optimal value. The cornerstone of our Multinoulli-SCG algorithm is an innovative continuous-relaxation framework named Multinoulli Extension(ME), which can effectively convert the discrete subset selection problem subject to partition constraints into a solvable continuous maximization focused on learning the optimal multinoulli priors across the concerned partition. In sharp contrast with the well-established multi-linear extension for submodular subset selection, a notable advantage of our proposed ME is its intrinsic capacity to provide a lossless rounding scheme for any set function. Furthermore, based on our proposed ME, we also present two novel online algorithms, namely, Multinoulli-OSCG and Multinoulli-OSGA, for the unexplored online subset selection problems over partition constraints.", "date": "2026-03-23", "categories": []} {"id": "2412.21071", "title": "Investigating layer-selective transfer learning of QAOA parameters for Max-Cut problem", "abstract": "The quantum approximate optimization algorithm (QAOA) is a variational quantum algorithm (VQA) ideal for noisy intermediate-scale quantum (NISQ) processors, and is highly successful in solving combinatorial optimization problems (COPs). It has been observed that the optimal parameters obtained from one instance of a COP can be transferred to another instance, resulting in generally good solutions for the latter. In this work, we propose a refinement scheme in which only a subset of QAOA layers is optimized following parameter transfer, with a focus on the Max-Cut problem. Our motivation is to reduce the complexity of the loss landscape when optimizing all the layers of high-depth QAOA circuits, as well as to reduce the optimization time. We investigate the potential hierarchical roles of different layers and analyze how the approximation ratio scales with increasing problem size. Our findings indicate that the selective layer optimization scheme offers a favorable trade-off between solution quality and computational time, and can be more beneficial than full optimization at a lower optimization time.", "date": "2024-12-30", "categories": []} {"id": "2506.08898", "title": "Preference-Driven Multi-Objective Combinatorial Optimization with Conditional Computation", "abstract": "Recent deep reinforcement learning methods have achieved remarkable success in solving multi-objective combinatorial optimization problems (MOCOPs) by decomposing them into multiple subproblems, each associated with a specific weight vector. However, these methods typically treat all subproblems equally and solve them using a single model, hindering the effective exploration of the solution space and thus leading to suboptimal performance. To overcome the limitation, we propose POCCO, a novel plug-and-play framework that enables adaptive selection of model structures for subproblems, which are subsequently optimized based on preference signals rather than explicit reward values. Specifically, we design a conditional computation block that routes subproblems to specialized neural architectures. Moreover, we propose a preference-driven optimization algorithm that learns pairwise preferences between winning and losing solutions. We evaluate the efficacy and versatility of POCCO by applying it to two state-of-the-art neural methods for MOCOPs. Experimental results across four classic MOCOP benchmarks demonstrate its significant superiority and strong generalization.", "date": "2025-06-10", "categories": []} {"id": "2506.16931", "title": "Multimodal Fused Learning for Solving the Generalized Traveling Salesman Problem in Robotic Task Planning", "abstract": "Effective and efficient task planning is essential for mobile robots, especially in applications like warehouse retrieval and environmental monitoring. These tasks often involve selecting one location from each of several target clusters, forming a Generalized Traveling Salesman Problem (GTSP) that remains challenging to solve both accurately and efficiently. To address this, we propose a Multimodal Fused Learning (MMFL) framework that leverages both graph and image-based representations to capture complementary aspects of the problem, and learns a policy capable of generating high-quality task planning schemes in real time. Specifically, we first introduce a coordinate-based image builder that transforms GTSP instances into spatially informative representations. We then design an adaptive resolution scaling strategy to enhance adaptability across different problem scales, and develop a multimodal fusion module with dedicated bottlenecks that enables effective integration of geometric and spatial features. Extensive experiments show that our MMFL approach significantly outperforms state-of-the-art methods across various GTSP instances while maintaining the computational efficiency required for real-time robotic applications. Physical robot tests further validate its practical effectiveness in real-world scenarios.", "date": "2025-06-20", "categories": []} {"id": "2602.21312", "title": "Precedence-Constrained Decision Trees and Coverings", "abstract": "This work considers a number of optimization problems and reductive relations between them. The two main problems we are interested in are the \\emph{Optimal Decision Tree} and \\emph{Set Cover}. We study these two fundamental tasks under precedence constraints, that is, if a test (or set) $X$ is a predecessor of $Y$, then in any feasible decision tree $X$ needs to be an ancestor of $Y$ (or respectively, if $Y$ is added to set cover, then so must be $X$). For the Optimal Decision Tree we consider two optimization criteria: worst case identification time (height of the tree) or the average identification time. Similarly, for the Set Cover we study two cost measures: the size of the cover or the average cover time.\n Our approach is to develop a number of algorithmic reductions, where an approximation algorithm for one problem provides an approximation for another via a black-box usage of a procedure for the former. En route we introduce other optimization problems either to complete the `reduction landscape' or because they hold the essence of combinatorial structure of our problems. The latter is brought by a problem of finding a maximum density precedence closed subfamily, where the density is defined as the ratio of the number of items the family covers to its size. By doing so we provide $\\mathcal{O}^*(\\sqrt{m})$-approximation algorithms for all of the aforementioned problems. The picture is complemented by a number of hardness reductions that provide $o(m^{1/12-\\epsilon})$-inapproximability results for the decision tree and covering problems. Besides giving a complete set of results for general precedence constraints, we also provide polylogarithmic approximation guarantees for two most typically studied and applicable precedence types, outforests and inforests. By providing corresponding hardness results, we show these results to be tight.", "date": "2026-02-24", "categories": []} {"id": "2603.19257", "title": "Constraint-aware Path Planning from Natural Language Instructions Using Large Language Models", "abstract": "Real-world path planning tasks typically involve multiple constraints beyond simple route optimization, such as the number of routes, maximum route length, depot locations, and task-specific requirements. Traditional approaches rely on dedicated formulations and algorithms for each problem variant, making them difficult to scale across diverse scenarios. In this work, we propose a flexible framework that leverages large language models (LLMs) to solve constrained path planning problems directly from natural language input. The core idea is to allow users to describe routing tasks conversationally, while enabling the LLM to interpret and solve the problem through solution verification and iterative refinement. The proposed method consists of two integrated components. For problem types that have been previously formulated and studied, the LLM first matches the input request to a known problem formulation in a library of pre-defined templates. For novel or unseen problem instances, the LLM autonomously infers a problem representation from the natural language description and constructs a suitable formulation in an in-context learning manner. In both cases, an iterative solution generation and verification process guides the LLM toward producing feasible and increasingly optimal solutions. Candidate solutions are compared and refined through multiple rounds of self-correction, inspired by genetic-algorithm-style refinement. We present the design, implementation, and evaluation of this LLM-based framework, demonstrating its capability to handle a variety of constrained path planning problems. This method provides a scalable and generalizable approach for solving real-world routing tasks with minimal human intervention, while enabling flexible problem specification through natural language.", "date": "2026-02-26", "categories": []} {"id": "2603.19318", "title": "Towards Solving Polynomial-Objective Integer Programming with Hypergraph Neural Networks", "abstract": "Complex real-world optimization problems often involve both discrete decisions and nonlinear relationships between variables. Many such problems can be modeled as polynomial-objective integer programs, encompassing cases with quadratic and higher-degree variable interactions. Nonlinearity makes them more challenging than their linear counterparts. In this paper, we propose a hypergraph neural network (HNN) based method to solve polynomial-objective integer programming (POIP). Besides presenting a high-degree-term-aware hypergraph representation to capture both high-degree information and variable-constraint interdependencies, we also propose a hypergraph neural network, which integrates convolution between variables and high-degree terms alongside convolution between variables and constraints, to predict solution values. Finally, a search process initialized from the predicted solutions is performed to further refine the results. Comprehensive experiments across a range of benchmarks demonstrate that our method consistently outperforms both existing learning-based approaches and state-of-the-art solvers, delivering superior solution quality with favorable efficiency. Note that our experiments involve both polynomial objectives and constraints, demonstrating our HNN's versatility for general POIP problems and highlighting its advancement over the existing literature.", "date": "2026-03-16", "categories": []} {"id": "2603.19074", "title": "CAMO: A Conditional Neural Solver for the Multi-objective Multiple Traveling Salesman Problem", "abstract": "Robotic systems often require a team of robots to collectively visit multiple targets while optimizing competing objectives, such as total travel cost and makespan. This setting can be formulated as the Multi-Objective Multiple Traveling Salesman Problem (MOMTSP). Although learning-based methods have shown strong performance on the single-agent TSP and multi-objective TSP variants, they rarely address the combined challenges of multi-agent coordination and multi-objective trade-offs, which introduce dual sources of complexity. To bridge this gap, we propose CAMO, a conditional neural solver for MOMTSP that generalizes across varying numbers of targets, agents, and preference vectors, and yields high-quality approximations to the Pareto front (PF). Specifically, CAMO consists of a conditional encoder to fuse preferences into instance representations, enabling explicit control over multi-objective trade-offs, and a collaborative decoder that coordinates all agents by alternating agent selection and node selection to construct multi-agent tours autoregressively. To further improve generalization, we train CAMO with a REINFORCE-based objective over a mixed distribution of problem sizes. Extensive experiments show that CAMO outperforms both neural and conventional heuristics, achieving a closer approximation of PFs. In addition, ablation results validate the contributions of CAMO's key components, and real-world tests on a mobile robot platform demonstrate its practical applicability.", "date": "2026-03-19", "categories": []} {"id": "2603.19163", "title": "cuGenOpt: A GPU-Accelerated General-Purpose Metaheuristic Framework for Combinatorial Optimization", "abstract": "Combinatorial optimization problems arise in logistics, scheduling, and resource allocation, yet existing approaches face a fundamental trade-off among generality, performance, and usability. We present cuGenOpt, a GPU-accelerated general-purpose metaheuristic framework that addresses all three dimensions simultaneously.\n At the engine level, cuGenOpt adopts a \"one block evolves one solution\" CUDA architecture with a unified encoding abstraction (permutation, binary, integer), a two-level adaptive operator selection mechanism, and hardware-aware resource management. At the extensibility level, a user-defined operator registration interface allows domain experts to inject problem-specific CUDA search operators. At the usability level, a JIT compilation pipeline exposes the framework as a pure-Python API, and an LLM-based modeling assistant converts natural-language problem descriptions into executable solver code.\n Experiments across five thematic suites on three GPU architectures (T4, V100, A800) show that cuGenOpt outperforms general MIP solvers by orders of magnitude, achieves competitive quality against specialized solvers on instances up to n=150, and attains 4.73% gap on TSP-442 within 30s. Twelve problem types spanning five encoding variants are solved to optimality. Framework-level optimizations cumulatively reduce pcb442 gap from 36% to 4.73% and boost VRPTW throughput by 75-81%.\n Code: https://github.com/L-yang-yang/cugenopt", "date": "2026-03-19", "categories": []} {"id": "2603.08447", "title": "Efficient Policy Learning with Hybrid Evaluation-Based Genetic Programming for Uncertain Agile Earth Observation Satellite Scheduling", "abstract": "The Uncertain Agile Earth Observation Satellite Scheduling Problem (UAEOSSP) is a novel combinatorial optimization problem and a practical engineering challenge that aligns with the current demands of space technology development. It incorporates uncertainties in profit, resource consumption, and visibility, which may render pre-planned schedules suboptimal or even infeasible. Genetic Programming Hyper-Heuristic (GPHH) shows promise for evolving interpretable scheduling policies; however, their simulation-based evaluation incurs high computational costs. Moreover, the design of the constructive method, denoted as Online Scheduling Algorithm (OSA), directly affects fitness assessment, resulting in evaluation-dependent local optima within the policy space. To address these issues, this paper proposes a Hybrid Evaluation-based Genetic Programming (HE-GP) for effectively solving UAEOSSP. A Hybrid Evaluation (HE) mechanism is integrated into the policy-driven OSA, combining exact and approximate filtering modes: exact mode ensures evaluation accuracy through elaborately designed constraint verification modules, while approximate mode reduces computational overhead via simplified logic. HE-GP dynamically switches between evaluation models based on real-time evolutionary state information. Experiments on 16 simulated instance sets demonstrate that HE-GP significantly outperforms handcrafted heuristics and single-evaluation based GPHH, achieving substantial reductions in computational cost while maintaining excellent scheduling performance across diverse scenarios. Specifically, the average training time of HE-GP was reduced by 17.77\\% compared to GP employing exclusively exact evaluation, while the optimal policy generated by HE-GP achieved the highest average ranks across all scenarios.", "date": "2026-03-09", "categories": []} {"id": "2507.03637", "title": "Large Language Models for Combinatorial Optimization: A Systematic Review", "abstract": "This systematic review explores the application of Large Language Models (LLMs) in Combinatorial Optimization (CO). We report our findings using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. We conduct a literature search via Scopus and Google Scholar, examining over 2,000 publications. We assess publications against four inclusion and four exclusion criteria related to their language, research focus, publication year, and type. Eventually, we select 103 studies. We classify these studies into semantic categories and topics to provide a comprehensive overview of the field, including the tasks performed by LLMs, the architectures of LLMs, the existing datasets specifically designed for evaluating LLMs in CO, and the field of application. Finally, we identify future directions for leveraging LLMs in this field.", "date": "2025-07-04", "categories": []} {"id": "2601.14485", "title": "Scalable Knee-Point Guided Activity Group Selection in Multi-Tree Genetic Programming for Dynamic Multi-Mode Project Scheduling", "abstract": "The dynamic multi-mode resource-constrained project scheduling problem is a challenging scheduling problem that requires making decisions on both the execution order of activities and their corresponding execution modes. Genetic programming has been widely applied as a hyper-heuristic to evolve priority rules that guide the selection of activity-mode pairs from the current eligible set. Recently, an activity group selection strategy has been proposed to select a subset of activities rather than a single activity at each decision point, allowing for more effective scheduling by considering the interdependence between activities. Although effective in small-scale instances, this strategy suffers from scalability issues when applied to larger problems. In this work, we enhance the scalability of the group selection strategy by introducing a knee-point-based selection mechanism to identify a promising subset of activities before evaluating their combinations. An activity ordering rule is first used to rank all eligible activity-mode pairs, followed by a knee point selection to find the promising pairs. Then, a group selection rule selects the best activity combination. We develop a multi-tree GP framework to evolve both types of rules simultaneously. Experimental results demonstrate that our approach scales well to large instances and outperforms GP with sequential decision-making in most scenarios.", "date": "2026-01-20", "categories": []} {"id": "2603.16286", "title": "Surrogate-Assisted Genetic Programming with Rank-Based Phenotypic Characterisation for Dynamic Multi-Mode Project Scheduling", "abstract": "The dynamic multi-mode resource-constrained project scheduling problem (DMRCPSP) is of practical importance, as it requires making real-time decisions under changing project states and resource availability. Genetic Programming (GP) has been shown to effectively evolve heuristic rules for such decision-making tasks; however, the evolutionary process typically relies on a large number of simulation-based fitness evaluations, resulting in high computational cost. Surrogate models offer a promising solution to reduce evaluation cost, but their application to GP requires problem-specific phenotypic characterisation (PC) schemes of heuristic rules. There is currently a lack of suitable PC schemes for GP applied to DMRCPSP.\n This paper proposes a rank-based PC scheme derived from heuristic-driven ordering of eligible activity-mode pairs and activity groups in decision situations. The resulting PC vectors enable a surrogate model to estimate the fitness of unevaluated GP individuals. Based on this scheme, a surrogate-assisted GP algorithm is developed. Experimental results demonstrate that the proposed surrogate-assisted GP can identify high-quality heuristic rules consistently earlier than the state-of-the-art GP approach for DMRCPSP, while introducing only marginal computational overhead. Further analyses demonstrate that the surrogate model provides useful guidance for offspring selection, leading to improved evolutionary efficiency.", "date": "2026-03-17", "categories": []} {"id": "2501.02942", "title": "Improved Approximation Algorithms for Orthogonally Constrained Problems Using Semidefinite Optimization", "abstract": "Building on the blueprint from Goemans and Williamson (1995) for the Max-Cut problem, we construct a polynomial-time approximation algorithm for orthogonally constrained quadratic optimization problems. First, we derive a semidefinite relaxation and propose a randomized rounding algorithm to generate feasible solutions from the relaxation. Second, we derive constant-factor approximation guarantees for our algorithm. When optimizing for $m$ orthonormal vectors in dimension $n$, we leverage strong duality and semidefinite complementary slackness to show that our algorithm achieves a $1/3$-approximation ratio. For any $m$ of the form $2^q$ for some integer $q$, we also construct an instance where the performance of our algorithm is exactly $(m+2)/(3m)$, which can be made arbitrarily close to $1/3$ by taking $m \\rightarrow + \\infty$, hence showing that our analysis is tight.", "date": "2025-01-06", "categories": []} {"id": "2503.16159", "title": "RRNCO: Towards Real-World Routing with Neural Combinatorial Optimization", "abstract": "The practical deployment of Neural Combinatorial Optimization (NCO) for Vehicle Routing Problems (VRPs) is hindered by a critical sim-to-real gap. This gap stems not only from training on oversimplified Euclidean data but also from node-based architectures incapable of handling the node-and-edge-based features with correlated asymmetric cost matrices, such as those for real-world distance and duration. We introduce RRNCO, a novel architecture specifically designed to address these complexities. RRNCO's novelty lies in two key innovations. First, its Adaptive Node Embedding (ANE) efficiently fuses spatial coordinates with real-world distance features using a learned contextual gating mechanism. Second, its Neural Adaptive Bias (NAB) is the first mechanism to jointly model asymmetric distance, duration, and directional angles, enabling it to capture complex, realistic routing constraints. Moreover, we introduce a new VRP benchmark grounded in real-world data crucial for bridging this sim-to-real gap, featuring asymmetric distance and duration matrices from 100 diverse cities, enabling the training and validation of NCO solvers on tasks that are more representative of practical settings. Experiments demonstrate that RRNCO achieves state-of-the-art performance on this benchmark, significantly advancing the practical applicability of neural solvers for real-world logistics. Our code, dataset, and pretrained models are available at https://github.com/ai4co/real-routing-nco.", "date": "2025-03-20", "categories": []} {"id": "2507.21386", "title": "Efficient Neural Combinatorial Optimization Solver for the Min-max Heterogeneous Capacitated Vehicle Routing Problem", "abstract": "Numerous Neural Combinatorial Optimization (NCO) solvers have been proposed to address Vehicle Routing Problems (VRPs). However, most of these solvers focus exclusively on single-vehicle VRP variants, overlooking the more realistic min-max Heterogeneous Capacitated Vehicle Routing Problem (MMHCVRP), which involves multiple vehicles. Existing MMHCVRP solvers typically select a vehicle and its next node to visit at each decoding step, but often make myopic decoding decisions and overlook key properties of MMHCVRP, including local topological relationships, vehicle permutation invariance, and node symmetry, resulting in suboptimal performance. To better address these limitations, we propose ECHO, an efficient NCO solver. First, ECHO exploits the proposed dual-modality node encoder to capture local topological relationships among nodes. Subsequently, to mitigate myopic decisions, ECHO employs the proposed Parameter-Free Cross-Attention mechanism to prioritize the vehicle selected in the preceding decoding step. Finally, leveraging vehicle permutation invariance and node symmetry, we introduce a tailored data augment strategy for MMHCVRP to stabilize the Reinforcement Learning training process. To assess the performance of ECHO, we conduct extensive experiments. The experimental results demonstrate that ECHO outperforms state-of-the-art NCO solvers across varying numbers of vehicles and nodes, and exhibits well-performing generalization across both scales and distribution patterns. Finally, ablation studies validate the effectiveness of all proposed methods.", "date": "2025-07-28", "categories": []} {"id": "2602.21761", "title": "Survey on Neural Routing Solvers", "abstract": "Neural routing solvers (NRSs) that leverage deep learning to tackle vehicle routing problems have demonstrated notable potential for practical applications. By learning implicit heuristic rules from data, NRSs replace the handcrafted counterparts in classic heuristic frameworks, thereby reducing reliance on costly manual design and trial-and-error adjustments. This survey makes two main contributions: (1) The heuristic nature of NRSs is highlighted, and existing NRSs are reviewed from the perspective of heuristics. A hierarchical taxonomy based on heuristic principles is further introduced. (2) A generalization-focused evaluation pipeline is proposed to address limitations of the conventional pipeline. Comparative benchmarking of representative NRSs across both pipelines uncovers a series of previously unreported gaps in current research.", "date": "2026-02-25", "categories": []} {"id": "2603.13344", "title": "DyACE: Dynamic Algorithm Co-evolution for Online Automated Heuristic Design with Large Language Model", "abstract": "The prevailing paradigm in Automated Heuristic Design (AHD) typically relies on the assumption that a single, fixed algorithm can effectively navigate the shifting dynamics of a combinatorial search. This static approach often proves inadequate for Perturbative Heuristics, where the optimal algorithm for escaping local optima depends heavily on the specific search phase. To address this limitation, we reformulate heuristic design as a Non-stationary Bi-level Control problem and introduce DyACE (Dynamic Algorithm Co-evolution). Distinct from standard open-loop solvers, DyACE use a Receding Horizon Control architecture to continuously co-evolve the heuristic logic alongside the solution population. A core element of this framework is the Look-Ahead Rollout Search, which queries the landscape geometry to extract Search Trajectory Features. This sensory feedback allows the Large Language Model (LLM) to function as a grounded meta-controller, prescribing phase-specific interventions tailored to the real-time search status. We validate DyACE on three representative combinatorial optimization benchmarks. The results demonstrate that our method significantly outperforms state-of-the-art static baselines, exhibiting superior scalability in high-dimensional search spaces. Furthermore, ablation studies confirm that dynamic adaptation fails without grounded perception, often performing worse than static algorithms. This indicates that DyACE's effectiveness stems from the causal alignment between the synthesized logic and the verified gradients of the optimization landscape.", "date": "2026-03-07", "categories": []} {"id": "2603.15218", "title": "Towards Foundation Models for Consensus Rank Aggregation", "abstract": "Aggregating a consensus ranking from multiple input rankings is a fundamental problem with applications in recommendation systems, search engines, job recruitment, and elections. Despite decades of research in consensus ranking aggregation, minimizing the Kemeny distance remains computationally intractable. Specifically, determining an optimal aggregation of rankings with respect to the Kemeny distance is an NP-hard problem, limiting its practical application to relatively small-scale instances. We propose the Kemeny Transformer, a novel Transformer-based algorithm trained via reinforcement learning to efficiently approximate the Kemeny optimal ranking. Experimental results demonstrate that our model outperforms classical majority-heuristic and Markov-chain approaches, achieving substantially faster inference than integer linear programming solvers. Our approach thus offers a practical, scalable alternative for real-world ranking-aggregation tasks.", "date": "2026-03-16", "categories": []} {"id": "2603.15541", "title": "Bridging Local and Global Knowledge: Cascaded Mixture-of-Experts Learning for Near-Shortest Path Routing", "abstract": "While deep learning models that leverage local features have demonstrated significant potential for near-optimal routing in dense Euclidean graphs, they struggle to generalize well in sparse networks where topological irregularities require broader structural awareness. To address this limitation, we train a Cascaded Mixture of Experts (Ca-MoE) to solve the all-pairs near-shortest path (APNSP) routing problem. Our Ca-MoE is a modular two-tier architecture that supports the decision-making for forwarder selection with lower-tier experts relying on local features and upper-tier experts relying on global features. It performs adaptive inference wherein the upper-tier experts are triggered only when the lower-tier ones do not suffice to achieve adequate decision quality. Computational efficiency is thus achieved by escalating model capacity only when necessitated by topological complexity, and parameter redundancy is avoided. Furthermore, we incorporate an online meta-learning strategy that facilitates independent expert fine-tuning and utilizes a stability-focused update mechanism to prevent catastrophic forgetting as new graph environments are encountered. Experimental evaluations demonstrate that Ca-MoE routing improves accuracy by up to 29.1% in sparse networks compared to single-expert baselines and maintains performance within 1%-6% of the theoretical upper bound across diverse graph densities.", "date": "2026-03-16", "categories": []} {"id": "2401.13883", "title": "Domain-Independent Dynamic Programming", "abstract": "For combinatorial optimization problems, model-based paradigms such as mixed-integer programming (MIP) and constraint programming (CP) aim to decouple modeling and solving a problem: the `holy grail' of declarative problem solving. We propose domain-independent dynamic programming (DIDP), a novel model-based paradigm based on dynamic programming (DP). While DP is not new, it has typically been implemented as a problem-specific method. We introduce Dynamic Programming Description Language (DyPDL), a formalism to define DP models based on a state transition system, inspired by artificial intelligence (AI) planning. we show that heuristic search algorithms can be used to solve DyPDL models and propose seven DIDP solvers. We experimentally compare our DIDP solvers with commercial MIP and CP solvers (solving MIP and CP models, respectively) on common benchmark instances of eleven combinatorial optimization problem classes. We show that DIDP outperforms MIP in nine problem classes, CP also in nine problem classes, and both MIP and CP in seven. DIDP also achieves superior performance to existing state-based solvers including domain-independent AI planners.", "date": "2024-01-25", "categories": []} {"id": "2601.03946", "title": "Provably Finding a Hidden Dense Submatrix among Many Planted Dense Submatrices via Convex Programming", "abstract": "We consider the densest submatrix problem, which seeks the submatrix of fixed size of a given binary matrix that contains the most nonzero entries. This problem is a natural generalization of fundamental problems in combinatorial optimization, e.g., the densest subgraph, maximum clique, and maximum edge biclique problems, and has wide application the study of complex networks. Much recent research has focused on the development of sufficient conditions for exact solution of the densest submatrix problem via convex relaxation. The vast majority of these sufficient conditions establish identification of the densest submatrix within a graph containing exactly one large dense submatrix hidden by noise. The assumptions of these underlying models are not observed in real-world networks, where the data may correspond to a matrix containing many dense submatrices of varying sizes.\n We extend and generalize these results to the more realistic setting where the input matrix may contain \\emph{many} large dense subgraphs. Specifically, we establish sufficient conditions under which we can expect to solve the densest submatrix problem in polynomial time for random input matrices sampled from a generalization of the stochastic block model. Moreover, we also provide sufficient conditions for perfect recovery under a deterministic adversarial. Numerical experiments involving randomly generated problem instances and real-world collaboration and communication networks are used empirically to verify the theoretical phase-transitions to perfect recovery given by these sufficient conditions.", "date": "2026-01-07", "categories": []} {"id": "2603.11456", "title": "UniHetCO: A Unified Heterogeneous Representation for Multi-Problem Learning in Unsupervised Neural Combinatorial Optimization", "abstract": "Unsupervised neural combinatorial optimization (NCO) offers an appealing alternative to supervised approaches by training learning-based solvers without ground-truth solutions, directly minimizing instance objectives and constraint violations. Yet for graph node subset-selection problems (e.g., Maximum Clique and Maximum Independent Set), existing unsupervised methods are typically specialized to a single problem class and rely on problem-specific surrogate losses, which hinders learning across classes within a unified framework. In this work, we propose UniHetCO, a unified heterogeneous graph representation for constrained quadratic programming-based combinatorial optimization that encodes problem structure, objective terms, and linear constraints in a single input. This formulation enables training a single model across multiple problem classes with a unified label-free objective. To improve stability under multi-problem learning, we employ a gradient-norm-based dynamic weighting scheme that alleviates gradient imbalance among classes. Experiments on multiple datasets and four constrained problem classes demonstrate competitive performance with state-of-the-art unsupervised NCO baselines, strong cross-problem adaptation potential, and effective warm starts for a commercial classical solver under tight time limits.", "date": "2026-03-12", "categories": []} {"id": "2603.12224", "title": "Portfolio of Solving Strategies in CEGAR-based Object Packing and Scheduling for Sequential 3D Printing", "abstract": "Computing power that used to be available only in supercomputers decades ago especially their parallelism is currently available in standard personal computer CPUs even in CPUs for mobile telephones. We show how to effectively utilize the computing power of modern multi-core personal computer CPU to solve the complex combinatorial problem of object arrangement and scheduling for sequential 3D printing. We achieved this by parallelizing the existing CEGAR-SEQ algorithm that solves the sequential object arrangement and scheduling by expressing it as a linear arithmetic formula which is then solved by a technique inspired by counterexample guided abstraction refinement (CEGAR). The original CEGAR-SEQ algorithm uses an object arrangement strategy that places objects towards the center of the printing plate. We propose alternative object arrangement strategies such as placing objects towards a corner of the printing plate and scheduling objects according to their height. Our parallelization is done at the high-level where we execute the CEGAR-SEQ algorithm in parallel with a portfolio of object arrangement strategies, an algorithm is called Porfolio-CEGAR-SEQ. Our experimental evaluation indicates that Porfolio-CEGAR-SEQ outperforms the original CEGAR-SEQ. When a batch of objects for multiple printing plates is scheduled, Portfolio-CEGAR-SEQ often uses fewer printing plates than CEGAR-SEQ.", "date": "2026-03-12", "categories": []} {"id": "2305.12292", "title": "Disjunctive Branch-and-Bound for Certifiably Optimal Low-Rank Matrix Completion", "abstract": "Low-rank matrix completion consists of computing a matrix of minimal complexity that recovers a given set of observations as accurately as possible. Unfortunately, existing methods for matrix completion are heuristics that, while highly scalable and often identifying high-quality solutions, do not provide an instance-wise certificate of optimality. We reexamine matrix completion with an optimality-oriented eye. We reformulate low-rank matrix completion problems as convex problems over the non-convex set of projection matrices and implement a disjunctive branch-and-bound scheme that solves them to certifiable optimality. Further, we derive a novel and often near-exact class of convex relaxations by decomposing a low-rank matrix as a sum of rank-one matrices and incentivizing that two-by-two minors in each rank-one matrix have determinant zero. In numerical experiments, our new convex relaxations decrease the optimality gap by two orders of magnitude compared to existing attempts, and our disjunctive branch-and-bound scheme solves $n \\times m$ rank-$k$ matrix completion problems to certifiable optimality or near optimality in hours for $\\max \\{m, n\\} \\leq 2500$ and $k \\leq 5$. Moreover, this reduction in the training error translates into an average $2\\%$--$50\\%$ reduction in the test set error compared with alternating minimization-based methods.", "date": "2023-05-20", "categories": []} {"id": "2502.12188", "title": "Boosting Cross-problem Generalization in Diffusion-Based Neural Combinatorial Solver via Inference Time Adaptation", "abstract": "Diffusion-based Neural Combinatorial Optimization (NCO) has demonstrated effectiveness in solving NP-complete (NPC) problems by learning discrete diffusion models for solution generation, eliminating hand-crafted domain knowledge. Despite their success, existing NCO methods face significant challenges in both cross-scale and cross-problem generalization, and high training costs compared to traditional solvers. While recent studies on diffusion models have introduced training-free guidance approaches that leverage pre-defined guidance functions for conditional generation, such methodologies have not been extensively explored in combinatorial optimization. To bridge this gap, we propose a training-free inference time adaptation framework (DIFU-Ada) that enables both the zero-shot cross-problem transfer and cross-scale generalization capabilities of diffusion-based NCO solvers without requiring additional training. We provide theoretical analysis that helps understanding the cross-problem transfer capability. Our experimental results demonstrate that a diffusion solver, trained exclusively on the Traveling Salesman Problem (TSP), can achieve competitive zero-shot transfer performance across different problem scales on TSP variants, such as Prize Collecting TSP (PCTSP) and the Orienteering Problem (OP), through inference time adaptation.", "date": "2025-02-15", "categories": []} {"id": "2602.23529", "title": "Active Value Querying to Minimize Additive Error in Subadditive Set Function Learning", "abstract": "Subadditive set functions play a pivotal role in computational economics (especially in combinatorial auctions), combinatorial optimization or artificial intelligence applications such as interpretable machine learning. However, specifying a set function requires assigning values to an exponentially large number of subsets in general, a task that is often resource-intensive in practice, particularly when the values derive from external sources such as retraining of machine learning models. A~simple omission of certain values introduces ambiguity that becomes even more significant when the incomplete set function has to be further optimized over. Motivated by the well-known result about inapproximability of subadditive functions using deterministic value queries with respect to a multiplicative error, we study a problem of approximating an unknown subadditive (or a subclass of thereof) set function with respect to an additive error -- i. e., we aim to efficiently close the distance between minimal and maximal completions. Our contributions are threefold: (i) a thorough exploration of minimal and maximal completions of different classes of set functions with missing values and an analysis of their resulting distance; (ii) the development of methods to minimize this distance over classes of set functions with a known prior, achieved by disclosing values of additional subsets in both offline and online manner; and (iii) empirical demonstrations of the algorithms' performance in practical scenarios.", "date": "2026-02-26", "categories": []} {"id": "2603.10053", "title": "Cluster-Aware Attention-Based Deep Reinforcement Learning for Pickup and Delivery Problems", "abstract": "The Pickup and Delivery Problem (PDP) is a fundamental and challenging variant of the Vehicle Routing Problem, characterized by tightly coupled pickup--delivery pairs, precedence constraints, and spatial layouts that often exhibit clustering. Existing deep reinforcement learning (DRL) approaches either model all nodes on a flat graph, relying on implicit learning to enforce constraints, or achieve strong performance through inference-time collaborative search at the cost of substantial latency. In this paper, we propose \\emph{CAADRL} (Cluster-Aware Attention-based Deep Reinforcement Learning), a DRL framework that explicitly exploits the multi-scale structure of PDP instances via cluster-aware encoding and hierarchical decoding. The encoder builds on a Transformer and combines global self-attention with intra-cluster attention over depot, pickup, and delivery nodes, producing embeddings that are both globally informative and locally role-aware. Based on these embeddings, we introduce a Dynamic Dual-Decoder with a learnable gate that balances intra-cluster routing and inter-cluster transitions at each step. The policy is trained end-to-end with a POMO-style policy gradient scheme using multiple symmetric rollouts per instance. Experiments on synthetic clustered and uniform PDP benchmarks show that CAADRL matches or improves upon strong state-of-the-art baselines on clustered instances and remains highly competitive on uniform instances, particularly as problem size increases. Crucially, our method achieves these results with substantially lower inference time than neural collaborative-search baselines, suggesting that explicitly modeling cluster structure provides an effective and efficient inductive bias for neural PDP solvers.", "date": "2026-03-09", "categories": []} {"id": "2505.16952", "title": "FrontierCO: Real-World and Large-Scale Evaluation of Machine Learning Solvers for Combinatorial Optimization", "abstract": "Machine learning (ML) has shown promise for tackling combinatorial optimization (CO), but much of the reported progress relies on small-scale, synthetic benchmarks that fail to capture real-world structure and scale. A core limitation is that ML methods are typically trained and evaluated on synthetic instance generators, leaving open how they perform on irregular, competition-grade, or industrial datasets. We present FrontierCO, a benchmark for evaluating ML-based CO solvers under real-world structure and extreme scale. FrontierCO spans eight CO problems, including routing, scheduling, facility location, and graph problems, with instances drawn from competitions and public repositories (e.g., DIMACS, TSPLib). Each task provides both easy sets (historically challenging but now solvable) and hard sets (open or computationally intensive), alongside standardized training/validation resources. Using FrontierCO, we evaluate 16 representative ML solvers--graph neural approaches, hybrid neural-symbolic methods, and LLM-based agents--against state-of-the-art classical solvers. We find a persistent performance gap that widens under structurally challenging and large instance sizes (e.g., TSP up to 10M nodes; MIS up to 8M), while also identifying cases where ML methods outperform classical solvers. By centering evaluation on real-world structure and orders-of-magnitude larger instances, FrontierCO provides a rigorous basis for advancing ML for CO. Our benchmark is available at https://huggingface.co/datasets/CO-Bench/FrontierCO.", "date": "2025-05-22", "categories": []} {"id": "2509.18057", "title": "Reinforced Generation of Combinatorial Structures: Hardness of Approximation", "abstract": "Can AI based methods help us make advances in complexity theory? We provide evidence towards answering this in the affirmative, using AlphaEvolve (an LLM code mutation agent) to obtain new results in three settings:\n a) We improve a recent result of Kunisky and Yu to obtain near-optimal upper and (conditional) lower bounds on certification algorithms for MAX-CUT and MAX-Independent Set on random 3- and 4-regular graphs. Our improved lower bounds are obtained by constructing nearly extremal Ramanujan graphs on as many as $163$ vertices, and our upper bounds are obtained via analytical arguments.\n b) We obtain new inapproximability results for MAX-4-CUT and MAX-3-CUT, proving that it is NP-hard to approximate them within factors of $0.987$ and $0.9649$ respectively, using AlphaEvolve to discover new gadget reductions. Our MAX-4-CUT result improves upon the SOTA of $0.9883$, and our MAX-3-CUT result improves on the current best gadget-based inapproximability result of $0.9853$, but falls short of the SOTA of $16/17$ that relies on a custom PCP (rather than a reduction from ``standard'' H{\\aa}stad-style PCPs).\n c) Inapproximability for the metric Traveling Salesman Problem (TSP): We show that it is NP-hard to approximate the minimum cost tour within a factor of $111/110$ using AlphaEvolve to discover a new gadget, thus improving the SOTA of $117/116$. Along the way, we provide new modular soundness and completeness arguments that can be of independent interest.\n A key technical challenge we faced: verifying a candidate construction produced by AlphaEvolve is costly (sometimes requiring time exponential in the size of the construction). We used AlphaEvolve itself to evolve the verification procedure to be faster (sometimes by $10,000\\times$ for our gadgets). Our results suggest that gadget based proofs would benefit from a pass through AI-based tools to obtain stronger results.", "date": "2025-09-22", "categories": []} {"id": "2602.10696", "title": "Robust Assortment Optimization from Observational Data", "abstract": "Assortment optimization is a fundamental challenge in modern retail and recommendation systems, where the goal is to select a subset of products that maximizes expected revenue under complex customer choice behaviors. While recent advances in data-driven methods have leveraged historical data to learn and optimize assortments, these approaches typically rely on strong assumptions -- namely, the stability of customer preferences and the correctness of the underlying choice models. However, such assumptions frequently break in real-world scenarios due to preference shifts and model misspecification, leading to poor generalization and revenue loss. Motivated by this limitation, we propose a robust framework for data-driven assortment optimization that accounts for potential distributional shifts in customer choice behavior. Our approach models potential preference shift from a nominal choice model that generates data and seeks to maximize worst-case expected revenue. We first establish the computational tractability of robust assortment planning when the nominal model is known, then advance to the data-driven setting, where we design statistically optimal algorithms that minimize the data requirements while maintaining robustness. Our theoretical analysis provides both upper bounds and matching lower bounds on the sample complexity, offering theoretical guarantees for robust generalization. Notably, we uncover and identify the notion of ``robust item-wise coverage'' as the minimal data requirement to enable sample-efficient robust assortment learning. Our work bridges the gap between robustness and statistical efficiency in assortment learning, contributing new insights and tools for reliable assortment optimization under uncertainty.", "date": "2026-02-11", "categories": []} {"id": "2503.11126", "title": "MUSS: Multilevel Subset Selection for Relevance and Diversity", "abstract": "The problem of relevant and diverse subset selection has a wide range of applications, including recommender systems and retrieval-augmented generation (RAG). For example, in recommender systems, one is interested in selecting relevant items, while providing a diversified recommendation. Constrained subset selection problem is NP-hard, and popular approaches such as Maximum Marginal Relevance (MMR) are based on greedy selection. Many real-world applications involve large data, but the original MMR work did not consider distributed selection. This limitation was later addressed by a method called DGDS which allows for a distributed setting using random data partitioning. Here, we exploit structure in the data to further improve both scalability and performance on the target application. We propose MUSS, a novel method that uses a multilevel approach to relevant and diverse selection. In a recommender system application, our method can not only improve the performance up to $4$ percent points in precision, but is also $20$ to $80$ times faster. Our method is also capable of outperforming baselines on RAG-based question answering accuracy. We present a novel theoretical approach for analyzing this type of problems, and show that our method achieves a constant factor approximation of the optimal objective. Moreover, our analysis also resulted in a $\\times 2$ tighter bound for DGDS compared to previously known bound.", "date": "2025-03-14", "categories": []} {"id": "2603.06777", "title": "HGT-Scheduler: Deep Reinforcement Learning for the Job Shop Scheduling Problem via Heterogeneous Graph Transformers", "abstract": "The Job Shop Scheduling Problem (JSSP) is commonly formulated as a disjunctive graph in which nodes represent operations and edges encode technological precedence constraints as well as machine-sharing conflicts. Most existing reinforcement learning approaches model this graph as homogeneous, merging job-precedence and machine-contention edges into a single relation type. Such a simplification overlooks the intrinsic heterogeneity of the problem structure and may lead to the loss of critical relational information. To address this limitation, we propose the Heterogeneous Graph Transformer (HGT)-Scheduler, a reinforcement learning framework that models the JSSP as a heterogeneous graph. The proposed architecture leverages a Heterogeneous Graph Transformer to capture type-specific relational patterns through edge-type-dependent attention mechanisms applied to precedence and contention relations. The scheduling policy is trained using Proximal Policy Optimization. The effectiveness of the proposed method is evaluated on the Fisher--Thompson benchmark instances. On the FT06 instance, the HGT-Scheduler achieves an optimality gap of 8.4\\%, statistically outperforming both an identical architecture that ignores edge types ($p = 0.011$) and a standard Graph Isomorphism Network baseline. On the larger FT10 instance, the approach demonstrates favorable scalability. However, under a 50,000-step training limit, the performance of heterogeneous and homogeneous graph models is comparable, suggesting that edge-type awareness requires longer training horizons for larger problem instances. Ablation analyses further indicate that a three-layer attention architecture provides the best performance. Overall, the results confirm that explicitly modeling distinct edge semantics improves the learning of effective scheduling policies.", "date": "2026-03-06", "categories": []} {"id": "2603.07020", "title": "RESCHED: Rethinking Flexible Job Shop Scheduling from a Transformer-based Architecture with Simplified States", "abstract": "Neural approaches to the Flexible Job Shop Scheduling Problem (FJSP), particularly those based on deep reinforcement learning (DRL), have gained growing attention in recent years. However, existing methods rely on complex feature-engineered state representations (i.e., often requiring more than 20 handcrafted features) and graph-biased neural architectures. To reduce modeling complexity and advance a more generalizable framework for FJSP, we introduce \\textsc{ReSched}, a minimalist DRL framework that rethinks both the scheduling formulation and model design. First, by revisiting the Markov Decision Process (MDP) formulation of FJSP, we condense the state space to just four essential features, eliminating historical dependencies through a subproblem-based perspective. Second, we employ Transformer blocks with dot-product attention, augmented by three lightweight but effective architectural modifications tailored to scheduling tasks. Extensive experiments show that \\textsc{ReSched} outperforms classical dispatching rules and state-of-the-art DRL methods on FJSP. Moreover, \\textsc{ReSched} also generalizes well to the Job Shop Scheduling Problem (JSSP) and the Flexible Flow Shop Scheduling Problem (FFSP), achieving competitive performance against neural baselines specifically designed for these variants.", "date": "2026-03-07", "categories": []} {"id": "2603.07373", "title": "Scheduling Parallel Optical Circuit Switches for AI Training", "abstract": "The rapid growth of AI training has dramatically increased datacenter traffic demand and energy consumption, which has motivated renewed interest in optical circuit switches (OCSes) as a high-bandwidth, energy-efficient alternative for AI fabrics. Deploying multiple parallel OCSes is a leading alternative. However, efficiently scheduling time-varying traffic matrices across parallel optical switches with non-negligible reconfiguration delays remains an open challenge.\n We consider the problem of scheduling a single AI traffic demand matrix $D$ over $s$ parallel OCSes while minimizing the makespan under reconfiguration delay $\\delta$. Our algorithm Spectra relies on a three-step approach: Decompose $D$ into a minimal set of weighted permutations; Schedule these permutations across parallel switches using load-aware assignment; then Equalize the imbalanced loads on the switches via controlled permutation splitting. Evaluated on realistic AI training workloads (GPT model and Qwen MoE expert routing) as well as standard benchmarks, Spectra vastly outperforms a baseline based on state-of-the-art algorithms, reducing schedule makespan by an average factor of $1.4\\times$ on GPT AI workloads, $1.9\\times$ on MoE AI workloads, and $2.4\\times$ on standard benchmarks. Further, the makespans achieved by Spectra consistently approach newly derived lower bounds.", "date": "2026-03-07", "categories": []} {"id": "2603.07568", "title": "Constraints Matrix Diffusion based Generative Neural Solver for Vehicle Routing Problems", "abstract": "Over the past decade, neural network solvers powered by generative artificial intelligence have garnered significant attention in the domain of vehicle routing problems (VRPs), owing to their exceptional computational efficiency and superior reasoning capabilities. In particular, autoregressive solvers integrated with reinforcement learning have emerged as a prominent trend. However, much of the existing work emphasizes large-scale generalization of neural approaches while neglecting the limited robustness of attention-based methods across heterogeneous distributions of problem parameters. Their improvements over heuristic search remain largely restricted to hand-curated, fixed-distribution benchmarks. Furthermore, these architectures tend to degrade significantly when node representations are highly similar or when tasks involve long decision horizons. To address the aforementioned limitations, we propose a novel fusion neural network framework that employs a discrete noise graph diffusion model to learn the underlying constraints of vehicle routing problems and generate a constraint assignment matrix. This matrix is subsequently integrated adaptively into the feature representation learning and decision process of the autoregressive solver, serving as a graph structure mask that facilitates the formation of solutions characterized by both global vision and local feature integration. To the best of our knowledge, this work represents the first comprehensive experimental investigation of neural network model solvers across a 378-combinatorial space spanning four distinct dimensions within the CVRPlib public dataset. Extensive experimental evaluations demonstrate that our proposed fusion model effectively captures and leverages problem constraints, achieving state-of-the-art performance across multiple benchmark datasets.", "date": "2026-03-08", "categories": []} {"id": "2603.07821", "title": "Column Generation for the Micro-Transit Zoning Problem", "abstract": "Along with the rapid development of new urban mobility options like ride-sharing over the past decade, on-demand micro-transit services stand out as a middle ground, bridging the gap between fixed-line mass transit and single-request ride-hailing, balancing ridership maximization and travel time minimization. Micro-transit adoption can have significant social impact. It improves urban sustainability, through lower energy consumption and reduced emissions, while enhancing equitable mobility access for disadvantaged communities, thanks to its lower vehicle miles per passenger, flexible schedules, and affordable pricing. However, effective operation of micro-transit services requires planning geo-fenced zones in advance, which involves solving a challenging combinatorial optimization problem. Existing approaches enumerate candidate zones first and selects a fixed number of optimal zones in the second step. In this paper, we generalize the Micro-Transit Zoning Problem (MZP) to allow a global budget rather than imposing a size limit for candidate zones. We also design a Column Generation (CG) framework to solve the problem and several pricing heuristics to accelerate computation. Extensive numerical experiments across major U.S. cities demonstrate that our approach produces higher-quality solutions more efficiently and scales better in the generalized setting.", "date": "2026-03-08", "categories": []} {"id": "2603.07891", "title": "A Lightweight Traffic Map for Efficient Anytime LaCAM*", "abstract": "Multi-Agent Path Finding (MAPF) aims to compute collision-free paths for multiple agents and has a wide range of practical applications. LaCAM*, an anytime configuration-based solver, currently represents the state of the art. Recent work has explored the use of guidance paths to steer LaCAM* toward configurations that avoid traffic congestion, thereby improving solution quality. However, existing approaches rely on Frank-Wolfe-style optimization that repeatedly invokes single-agent search before executing LaCAM*, resulting in substantial computational overhead for large-scale problems. Moreover, the guidance path is static and primarily beneficial for finding the first solution in LaCAM*. To address these limitations, we propose a new approach that leverages LaCAM*'s ability to construct a dynamic, lightweight traffic map during its search. Experimental results demonstrate that our method achieves higher solution quality than state-of-the-art guidance-path approaches across two MAPF variants.", "date": "2026-03-09", "categories": []} {"id": "2603.08188", "title": "Sequential Service Region Design with Capacity-Constrained Investment and Spillover Effect", "abstract": "Service region design determines the geographic coverage of service networks, shaping long-term operational performance. Capital and operational constraints preclude simultaneous large-scale deployment, requiring expansion to proceed sequentially. The resulting challenge is to determine when and where to invest under demand uncertainty, balancing intertemporal trade-offs between early and delayed investment and accounting for network effects whereby each deployment reshapes future demand through inter-regional connectivity. This study addresses a sequential service region design (SSRD) problem incorporating two practical yet underexplored factors: a $k$-region constraint that limits the number of regions investable per period and a stochastic spillover effect linking investment decisions to demand evolution. The resulting problem requires sequencing regional portfolios under uncertainty, leading to a combinatorial explosion in feasible investment sequences. To address this challenge, we propose a solution framework that integrates real options analysis (ROA) with a Transformer-based Proximal Policy Optimization (TPPO) algorithm. ROA evaluates the intertemporal option value of investment sequences, while TPPO learns sequential policies that directly generate high option-value sequences without exhaustive enumeration. Numerical experiments on realistic multi-region settings demonstrate that TPPO converges faster than benchmark DRL methods and consistently identifies sequences with superior option value. Case studies and sensitivity analyses further confirm robustness and provide insights on investment concurrency, regional prioritization, and the increasing benefits of adaptive expansion via our approach under stronger spillovers and dynamic market conditions.", "date": "2026-03-09", "categories": []} {"id": "2603.08283", "title": "PolyFormer: learning efficient reformulations for scalable optimization under complex physical constraints", "abstract": "Real-world optimization problems are often constrained by complex physical laws that limit computational scalability. These constraints are inherently tied to complex regions, and thus learning models that incorporate physical and geometric knowledge, i.e., physics-informed machine learning (PIML), offer a promising pathway for efficient solution. Here, we introduce PolyFormer, which opens a new direction for PIML in prescriptive optimization tasks, where physical and geometric knowledge is not merely used to regularize learning models, but to simplify the problems themselves. PolyFormer captures geometric structures behind constraints and transforms them into efficient polytopic reformulations, thereby decoupling problem complexity from solution difficulty and enabling off-the-shelf optimization solvers to efficiently produce feasible solutions with acceptable optimality loss. Through evaluations across three important problems (large-scale resource aggregation, network-constrained optimization, and optimization under uncertainty), PolyFormer achieves computational speedups up to 6,400-fold and memory reductions up to 99.87%, while maintaining solution quality competitive with or superior to state-of-the-art methods. These results demonstrate that PolyFormer provides an efficient and reliable solution for scalable constrained optimization, expanding the scope of PIML to prescriptive tasks in scientific discovery and engineering applications.", "date": "2026-03-09", "categories": []} {"id": "2603.05579", "title": "A Novel Hybrid Heuristic-Reinforcement Learning Optimization Approach for a Class of Railcar Shunting Problems", "abstract": "Railcar shunting is a core planning task in freight railyards, where yard planners need to disassemble and reassemble groups of railcars to form outbound trains. Classification tracks with access from one side only can be considered as stack structures, where railcars are added and removed from only one end, leading to a last-in-first-out (LIFO) retrieval order. In contrast, two-sided tracks function like queue structures, allowing railcars to be added from one end and removed from the opposite end, following a first-in-first-out (FIFO) order. We consider a problem requiring assembly of multiple outbound trains using two locomotives in a railyard with two-sided classification track access. To address this combinatorially challenging problem class, we decompose the problem into two subproblems, each with one-sided classification track access and a locomotive on each side. We present a novel Hybrid Heuristic-Reinforcement Learning (HHRL) framework that integrates railway-specific heuristic solution approaches with a reinforcement learning method, specifically Q-learning. The proposed framework leverages methods to decrease the state-action space and guide exploration during reinforcement learning. The results of a series of numerical experiments demonstrate the efficiency and quality of the HHRL algorithm in both one-sided access, single-locomotive problems and two-sided access, two-locomotive problems.", "date": "2026-03-05", "categories": []} {"id": "2603.06260", "title": "Learning to Solve Orienteering Problem with Time Windows and Variable Profits", "abstract": "The orienteering problem with time windows and variable profits (OPTWVP) is common in many real-world applications and involves continuous time variables. Current approaches fail to develop an efficient solver for this orienteering problem variant with discrete and continuous variables. In this paper, we propose a learning-based two-stage DEcoupled discrete-Continuous optimization with Service-time-guided Trajectory (DeCoST), which aims to effectively decouple the discrete and continuous decision variables in the OPTWVP problem, while enabling efficient and learnable coordination between them. In the first stage, a parallel decoding structure is employed to predict the path and the initial service time allocation. The second stage optimizes the service times through a linear programming (LP) formulation and provides a long-horizon learning of structure estimation. We rigorously prove the global optimality of the second-stage solution. Experiments on OPTWVP instances demonstrate that DeCoST outperforms both state-of-the-art constructive solvers and the latest meta-heuristic algorithms in terms of solution quality and computational efficiency, achieving up to 6.6x inference speedup on instances with fewer than 500 nodes. Moreover, the proposed framework is compatible with various constructive solvers and consistently enhances the solution quality for OPTWVP.", "date": "2026-03-06", "categories": []} {"id": "2404.03759", "title": "Localized Distributional Robustness in Submodular Multi-Task Subset Selection", "abstract": "In this work, we treat the problem of multi-task submodular optimization from the perspective of local distributional robustness within the neighborhood of a reference distribution which assigns an importance score to each task. We initially propose to introduce a relative-entropy regularization term to the standard multi-task objective. We then demonstrate through duality that this novel formulation itself is equivalent to the maximization of a monotone increasing function composed with a submodular function, which may be efficiently carried out through standard greedy selection methods. This approach bridges the existing gap in the optimization of performance-robustness trade-offs in multi-task subset selection. To numerically validate our theoretical results, we test the proposed method in two different settings, one on the selection of satellites in low Earth orbit constellations in the context of a sensor selection problem involving weak-submodular functions, and the other on an image summarization task using neural networks involving submodular functions. Our method is compared with two other algorithms focused on optimizing the performance of the worst-case task, and on directly optimizing the performance on the reference distribution itself. We conclude that our novel formulation produces a solution that is locally distributional robust, and computationally inexpensive.", "date": "2024-04-04", "categories": []} {"id": "2601.01832", "title": "Yukthi Opus: A Multi-Chain Hybrid Metaheuristic for Large-Scale NP-Hard Optimization", "abstract": "We present Yukthi Opus (YO), a multi-chain hybrid metaheuristic designed for NP-hard optimization under explicit evaluation budget constraints. YO integrates three complementary mechanisms in a structured two-phase architecture: Markov Chain Monte Carlo (MCMC) for global exploration, greedy local search for exploitation, and simulated annealing with adaptive reheating to enable controlled escape from local minima. A dedicated burn-in phase allocates evaluations to probabilistic exploration, after which a hybrid optimization loop refines promising candidates. YO further incorporates a spatial blacklist mechanism to avoid repeated evaluation of poor regions and a multi-chain execution strategy to improve robustness and reduce sensitivity to initialization.\n We evaluate YO on three benchmarks: the Rastrigin function (5D) with ablation studies, the Traveling Salesman Problem with 50 to 200 cities, and the Rosenbrock function (5D) with comparisons against established optimizers including CMA-ES, Bayesian optimization, and accelerated particle swarm optimization. Results show that MCMC exploration and greedy refinement are critical for solution quality, while simulated annealing and multi-chain execution primarily improve stability and variance reduction. Overall, YO achieves competitive performance on large and multimodal problems while maintaining predictable evaluation budgets, making it suitable for expensive black-box optimization settings.", "date": "2026-01-05", "categories": []} {"id": "2603.03388", "title": "RADAR: Learning to Route with Asymmetry-aware DistAnce Representations", "abstract": "Recent neural solvers have achieved strong performance on vehicle routing problems (VRPs), yet they mainly assume symmetric Euclidean distances, restricting applicability to real-world scenarios. A core challenge is encoding the relational features in asymmetric distance matrices of VRPs. Early attempts directly encoded these matrices but often failed to produce compact embeddings and generalized poorly at scale. In this paper, we propose RADAR, a scalable neural framework that augments existing neural VRP solvers with the ability to handle asymmetric inputs. RADAR addresses asymmetry from both static and dynamic perspectives. It leverages Singular Value Decomposition (SVD) on the asymmetric distance matrix to initialize compact and generalizable embeddings that inherently encode the static asymmetry in the inbound and outbound costs of each node. To further model dynamic asymmetry in embedding interactions during encoding, it replaces the standard softmax with Sinkhorn normalization that imposes joint row and column distance awareness in attention weights. Extensive experiments on synthetic and real-world benchmarks across various VRPs show that RADAR outperforms strong baselines on both in-distribution and out-of-distribution instances, demonstrating robust generalization and superior performance in solving asymmetric VRPs.", "date": "2026-03-03", "categories": []} {"id": "2301.00061", "title": "A Global Optimization Algorithm for K-Center Clustering of One Billion Samples", "abstract": "This paper presents a practical global optimization algorithm for the K-center clustering problem, which aims to select K samples as the cluster centers to minimize the maximum within-cluster distance. This algorithm is based on a reduced-space branch and bound scheme and guarantees convergence to the global optimum in a finite number of steps by only branching on the regions of centers. To improve efficiency, we have designed a two-stage decomposable lower bound, the solution of which can be derived in a closed form. In addition, we also propose several acceleration techniques to narrow down the region of centers, including bounds tightening, sample reduction, and parallelization. Extensive studies on synthetic and real-world datasets have demonstrated that our algorithm can solve the K-center problems to global optimal within 4 hours for ten million samples in the serial mode and one billion samples in the parallel mode. Moreover, compared with the state-of-the-art heuristic methods, the global optimum obtained by our algorithm can averagely reduce the objective function by 25.8% on all the synthetic and real-world datasets.", "date": "2022-12-30", "categories": []} {"id": "2603.02462", "title": "Can Computational Reducibility Lead to Transferable Models for Graph Combinatorial Optimization?", "abstract": "A key challenge in deriving unified neural solvers for combinatorial optimization (CO) is efficient generalization of models between a given set of tasks to new tasks not used during the initial training process. To address it, we first establish a new model, which uses a GCON module as a form of expressive message passing together with energy-based unsupervised loss functions. This model achieves high performance (often comparable with state-of-the-art results) across multiple CO tasks when trained individually on each task. We then leverage knowledge from the computational reducibility literature to propose pretraining and fine-tuning strategies that transfer effectively (a) between MVC, MIS and MaxClique, and (b) in a multi-task learning setting that additionally incorporates MaxCut, MDS and graph coloring. Additionally, in a leave-one-out, multi-task learning setting, we observe that pretraining on all but one task almost always leads to faster convergence on the remaining task when fine-tuning while avoiding negative transfer. Our findings indicate that learning common representations across multiple graph CO problems is viable through the use of expressive message passing coupled with pretraining strategies that are informed by the polynomial reduction literature, thereby taking an important step towards enabling the development of foundational models for neural CO. We provide an open-source implementation of our work at https://github.com/semihcanturk/COPT-MT .", "date": "2026-03-02", "categories": []} {"id": "2603.02846", "title": "Learning Memory-Enhanced Improvement Heuristics for Flexible Job Shop Scheduling", "abstract": "The rise of smart manufacturing under Industry 4.0 introduces mass customization and dynamic production, demanding more advanced and flexible scheduling techniques. The flexible job-shop scheduling problem (FJSP) has attracted significant attention due to its complex constraints and strong alignment with real-world production scenarios. Current deep reinforcement learning (DRL)-based approaches to FJSP predominantly employ constructive methods. While effective, they often fall short of reaching (near-)optimal solutions. In contrast, improvement-based methods iteratively explore the neighborhood of initial solutions and are more effective in approaching optimality. However, the flexible machine allocation in FJSP poses significant challenges to the application of this framework, including accurate state representation, effective policy learning, and efficient search strategies. To address these challenges, this paper proposes a Memory-enhanced Improvement Search framework with heterogeneous graph representation--MIStar. It employs a novel heterogeneous disjunctive graph that explicitly models the operation sequences on machines to accurately represent scheduling solutions. Moreover, a memoryenhanced heterogeneous graph neural network (MHGNN) is designed for feature extraction, leveraging historical trajectories to enhance the decision-making capability of the policy network. Finally, a parallel greedy search strategy is adopted to explore the solution space, enabling superior solutions with fewer iterations. Extensive experiments on synthetic data and public benchmarks demonstrate that MIStar significantly outperforms both traditional handcrafted improvement heuristics and state-of-the-art DRL-based constructive methods.", "date": "2026-03-03", "categories": []} {"id": "2603.03230", "title": "SynthCharge: An Electric Vehicle Routing Instance Generator with Feasibility Screening to Enable Learning-Based Optimization and Benchmarking", "abstract": "The electric vehicle routing problem with time windows (EVRPTW) extends the classical VRPTW by introducing battery capacity constraints and charging station decisions. Existing benchmark datasets are often static and lack verifiable feasibility, which restricts reproducible evaluation of learning-based routing models. We introduce SynthCharge, a parametric generator that produces diverse, feasibility-screened EVRPTW instances across varying spatiotemporal configurations and scalable customer counts. While SynthCharge can currently generate large-scale instances of up to 500 customers, we focus our experiments on sizes ranging from 5 to 100 customers. Unlike static benchmark suites, SynthCharge integrates instance geometry with adaptive energy capacity scaling and range-aware charging station placement. To guarantee structural validity, the generator systematically filters out unsolvable instances through a fast feasibility screening process. Ultimately, SynthCharge provides the dynamic benchmarking infrastructure needed to systematically evaluate the robustness of emerging neural routing and data-driven approaches.", "date": "2026-03-03", "categories": []} {"id": "2504.12419", "title": "Standardization of Multi-Objective QUBOs", "abstract": "Multi-objective optimization involving Quadratic Unconstrained Binary Optimization (QUBO) problems arises in various domains. A fundamental challenge in this context is the effective balancing of multiple objectives, each potentially operating on very different scales. This imbalance introduces complications such as the selection of appropriate weights when scalarizing multiple objectives into a single objective function. In this paper, we propose a novel technique for scaling QUBO objectives that uses an exact computation of the variance of each individual QUBO objective. By scaling each objective to have unit variance, we align all objectives onto a common scale, thereby allowing for more balanced solutions to be found when scalarizing the objectives with equal weights, as well as potentially assisting in the search or choice of weights during scalarization. Finally, we demonstrate its advantages through empirical evaluations on various multi-objective optimization problems. Our results are noteworthy since manually selecting scalarization weights is cumbersome, and reliable, efficient solutions are scarce.", "date": "2025-04-16", "categories": []} {"id": "2505.03776", "title": "PAPN: Proximity Attention Encoder and Pointer Network Decoder for Parcel Pickup Route Prediction", "abstract": "Optimization of the last-mile delivery and first-mile pickup of parcels is integral to the logistics optimization pipeline as it entails both cost and resource efficiency and a heightened service quality. Such optimization requires accurate route and time prediction systems to adapt to different scenarios in advance. This work tackles the first building block, namely route prediction. The novel Proximity Attention (PA) mechanism is coupled to a Pointer Network (PN) decoder to leverage the underlying connections between the different visitable pickup positions at each timestep of the parcel pickup process. This local attention is coupled with global context computing via a multi-head attention transformer encoder. Both attentions are then mixed for complete and comprehensive modeling of the problems. PA is also used in the decoding process to skew predictions towards the locations with the highest visit likeliness, thus using inter-connectivity of nodes for next-location prediction. This method is trained, validated and tested on a large industry-level dataset of real-world, last-mile delivery and first-mile pickup named LaDE (2024). This approach outperforms all state-of-the-art supervised methods in terms of most metrics used for benchmarking on this dataset while still being competitive with the best-performing reinforcement learning framework named DRL4Route (2023).", "date": "2025-04-30", "categories": []} {"id": "2505.17108", "title": "REMS: a unified solution representation, problem modeling and metaheuristic algorithm design for general combinatorial optimization problems", "abstract": "Combinatorial optimization problems (COPs) with discrete variables and finite search space are critical across numerous fields, and solving them in metaheuristic algorithms is popular. However, addressing a specific COP typically requires developing a tailored and handcrafted algorithm. Even minor adjustments, such as constraint changes, may necessitate algorithm redevelopment. Therefore, establishing a framework for formulating diverse COPs into a unified paradigm and designing reusable metaheuristic algorithms is valuable. A COP can be typically viewed as the process of giving resources to perform specific tasks, subjecting to given constraints. Motivated by this, a resource-centered modeling and solving framework (REMS) is introduced for the first time. We first extract and define resources and tasks from a COP. Subsequently, given predetermined resources, the solution structure is unified as assigning tasks to resources, from which variables, objectives, and constraints can be derived and a problem model is constructed. To solve the modeled COPs, several fundamental operators are designed based on the unified solution structure, including the initial solution, neighborhood structure, destruction and repair, crossover, and ranking. These operators enable the development of various metaheuristic algorithms. Specially, 4 single-point-based algorithms and 1 population-based algorithm are configured herein. Experiments on 10 COPs, covering routing, location, loading, assignment, scheduling, and graph coloring problems, show that REMS can model these COPs within the unified paradigm and effectively solve them with the designed metaheuristic algorithms. Furthermore, REMS is more competitive than GUROBI and SCIP in tackling large-scale instances and complex COPs, and outperforms OR-TOOLS on several challenging COPs.", "date": "2025-05-21", "categories": []} {"id": "2509.23465", "title": "ViTSP: A Vision Language Models Guided Framework for Solving Large-Scale Traveling Salesman Problems", "abstract": "Solving the Traveling Salesman Problem (TSP) is NP-hard yet fundamental for a wide range of real-world applications. Classical exact methods face challenges in scaling, and heuristic methods often require domain-specific parameter calibration. While learning-based approaches have shown promise, they suffer from poor generalization and limited scalability due to fixed training data. This work proposes ViTSP, a novel framework that leverages pre-trained vision language models (VLMs) to visually guide the solution process for large-scale TSPs. The VLMs function to identify promising small-scale subproblems from a visualized TSP instance, which are then efficiently optimized using an off-the-shelf solver to improve the global solution. ViTSP bypasses the dedicated model training at the user end while maintaining effectiveness across diverse instances. Experiments on real-world TSP instances ranging from 1k to 88k nodes demonstrate that ViTSP consistently achieves solutions with average optimality gaps of 0.24%, outperforming existing learning-based methods. Under the same runtime budget, it surpasses the best-performing heuristic solver, LKH-3, by reducing its gaps by 3.57% to 100%, particularly on very-large-scale instances with more than 10k nodes. Our framework offers a new perspective in hybridizing pre-trained generative models and operations research solvers in solving combinatorial optimization problems. The framework holds potential for integration into more complex real-world logistics systems. The code is available at https://github.itap.purdue.edu/uSMART/ViTSP_ICLR2026.", "date": "2025-09-27", "categories": []} {"id": "2601.06502", "title": "DRAGON: LLM-Driven Decomposition and Reconstruction Agents for Large-Scale Combinatorial Optimization", "abstract": "Large Language Models (LLMs) have recently shown promise in addressing combinatorial optimization problems (COPs) through prompt-based strategies. However, their scalability and generalization remain limited, and their effectiveness diminishes as problem size increases, particularly in routing problems involving more than 30 nodes. We propose DRAGON, which stands for Decomposition and Reconstruction Agents Guided OptimizatioN, a novel framework that combines the strengths of metaheuristic design and LLM reasoning. Starting from an initial global solution, DRAGON autonomously identifies regions with high optimization potential and strategically decompose large-scale COPs into manageable subproblems. Each subproblem is then reformulated as a concise, localized optimization task and solved through targeted LLM prompting guided by accumulated experiences. Finally, the locally optimized solutions are systematically reintegrated into the original global context to yield a significantly improved overall outcome. By continuously interacting with the optimization environment and leveraging an adaptive experience memory, the agents iteratively learn from feedback, effectively coupling symbolic reasoning with heuristic search. Empirical results show that, unlike existing LLM-based solvers limited to small-scale instances, DRAGON consistently produces feasible solutions on TSPLIB, CVRPLIB, and Weibull-5k bin packing benchmarks, and achieves near-optimal results (0.16% gap) on knapsack problems with over 3M variables. This work shows the potential of feedback-driven language agents as a new paradigm for generalizable and interpretable large-scale optimization.", "date": "2026-01-10", "categories": []} {"id": "2602.08542", "title": "Incremental (k, z)-Clustering on Graphs", "abstract": "Given a weighted undirected graph, a number of clusters $k$, and an exponent $z$, the goal in the $(k, z)$-clustering problem on graphs is to select $k$ vertices as centers that minimize the sum of the distances raised to the power $z$ of each vertex to its closest center. In the dynamic setting, the graph is subject to adversarial edge updates, and the goal is to maintain explicitly an exact $(k, z)$-clustering solution in the induced shortest-path metric.\n While efficient dynamic $k$-center approximation algorithms on graphs exist [Cruciani et al. SODA 2024], to the best of our knowledge, no prior work provides similar results for the dynamic $(k,z)$-clustering problem. As the main result of this paper, we develop a randomized incremental $(k, z)$-clustering algorithm that maintains with high probability a constant-factor approximation in a graph undergoing edge insertions with a total update time of $\\tilde O(k m^{1+o(1)}+ k^{1+\\frac{1}{\\lambda}} m)$, where $\\lambda \\geq 1$ is an arbitrary fixed constant. Our incremental algorithm consists of two stages. In the first stage, we maintain a constant-factor bicriteria approximate solution of size $\\tilde{O}(k)$ with a total update time of $m^{1+o(1)}$ over all adversarial edge insertions. This first stage is an intricate adaptation of the bicriteria approximation algorithm by Mettu and Plaxton [Machine Learning 2004] to incremental graphs. One of our key technical results is that the radii in their algorithm can be assumed to be non-decreasing while the approximation ratio remains constant, a property that may be of independent interest.\n In the second stage, we maintain a constant-factor approximate $(k,z)$-clustering solution on a dynamic weighted instance induced by the bicriteria approximate solution. For this subproblem, we employ a dynamic spanner algorithm together with a static $(k,z)$-clustering algorithm.", "date": "2026-02-09", "categories": []} {"id": "2603.00174", "title": "Automated Discovery of Improved Constant Weight Binary Codes", "abstract": "A constant weight binary code consists of $n$-bit binary codewords, each with exactly $w$ bits equal to 1, such that any two codewords are at least Hamming distance $d$ apart. $A(n,d,w)$ is the maximum size of a constant weight binary code with parameters $n,d,w$. We establish improved lower bounds on $A(n,d,w)$ by constructing new larger codes, for 24 values of $(n,d,w)$ with $6 \\leq d \\leq 18$ and $18 \\leq n \\leq 35$. The improved lower bounds come from two strategies. The first is a tabu search that operates at the level of bit swaps. The second is a novel greedy heuristic that repeatedly chooses the candidate codeword that maximizes a randomly-scored histogram of distances to previously-added codewords. These strategies were proposed by CPro1, an automated protocol that generates, implements, and tests diverse strategies for combinatorial constructions.", "date": "2026-02-26", "categories": []} {"id": "2603.00430", "title": "Efficient Decoder Scaling Strategy for Neural Routing Solvers", "abstract": "Construction-based neural routing solvers, typically composed of an encoder and a decoder, have emerged as a promising approach for solving vehicle routing problems. While recent studies suggest that shifting parameters from the encoder to the decoder enhances performance, most works restrict the decoder size to 1-3M parameters, leaving the effects of scaling largely unexplored. To address this gap, we conduct a systematic study comparing two distinct strategies: scaling depth versus scaling width. We synthesize these strategies to construct a suite of 12 model configurations, spanning a parameter range from 1M to ~150M, and extensively evaluate their scaling behaviors across three critical dimensions: parameter efficiency, data efficiency, and compute efficiency. Our empirical results reveal that parameter count is insufficient to accurately predict the model performance, highlighting the critical and distinct roles of model depth (layer count) and width (embedding dimension). Crucially, we demonstrate that scaling depth yields superior performance gains to scaling width. Based on these findings, we provide and experimentally validate a set of design principles for the efficient allocation of parameters and compute resources to enhance the model performance.", "date": "2026-02-28", "categories": []} {"id": "2603.01667", "title": "Chain-of-Context Learning: Dynamic Constraint Understanding for Multi-Task VRPs", "abstract": "Multi-task Vehicle Routing Problems (VRPs) aim to minimize routing costs while satisfying diverse constraints. Existing solvers typically adopt a unified reinforcement learning (RL) framework to learn generalizable patterns across tasks. However, they often overlook the constraint and node dynamics during the decision process, making the model fail to accurately react to the current context. To address this limitation, we propose Chain-of-Context Learning (CCL), a novel framework that progressively captures the evolving context to guide fine-grained node adaptation. Specifically, CCL constructs step-wise contextual information via a Relevance-Guided Context Reformulation (RGCR) module, which adaptively prioritizes salient constraints. This context then guides node updates through a Trajectory-Shared Node Re-embedding (TSNR) module, which aggregates shared node features from all trajectories' contexts and uses them to update inputs for the next step. By modeling evolving preferences of the RL agent, CCL captures step-by-step dependencies in sequential decision-making. We evaluate CCL on 48 diverse VRP variants, including 16 in-distribution and 32 out-of-distribution (with unseen constraints) tasks. Experimental results show that CCL performs favorably against the state-of-the-art baselines, achieving the best performance on all in-distribution tasks and the majority of out-of-distribution tasks.", "date": "2026-03-02", "categories": []} {"id": "2603.01786", "title": "Learning Shortest Paths with Generative Flow Networks", "abstract": "In this paper, we present a novel learning framework for finding shortest paths in graphs utilizing Generative Flow Networks (GFlowNets). First, we examine theoretical properties of GFlowNets in non-acyclic environments in relation to shortest paths. We prove that, if the total flow is minimized, forward and backward policies traverse the environment graph exclusively along shortest paths between the initial and terminal states. Building on this result, we show that the pathfinding problem in an arbitrary graph can be solved by training a non-acyclic GFlowNet with flow regularization. We experimentally demonstrate the performance of our method in pathfinding in permutation environments and in solving Rubik's Cubes. For the latter problem, our approach shows competitive results with state-of-the-art machine learning approaches designed specifically for this task in terms of the solution length, while requiring smaller search budget at test-time.", "date": "2026-03-02", "categories": []} {"id": "2402.01446", "title": "Guidance Graph Optimization for Lifelong Multi-Agent Path Finding", "abstract": "We study how to use guidance to improve the throughput of lifelong Multi-Agent Path Finding (MAPF). Previous studies have demonstrated that, while incorporating guidance, such as highways, can accelerate MAPF algorithms, this often results in a trade-off with solution quality. In addition, how to generate good guidance automatically remains largely unexplored, with current methods falling short of surpassing manually designed ones. In this work, we introduce the guidance graph as a versatile representation of guidance for lifelong MAPF, framing Guidance Graph Optimization as the task of optimizing its edge weights. We present two GGO algorithms to automatically generate guidance for arbitrary lifelong MAPF algorithms and maps. The first method directly optimizes edge weights, while the second method optimizes an update model capable of generating edge weights. Empirically, we show that (1) our guidance graphs improve the throughput of three representative lifelong MAPF algorithms in eight benchmark maps, and (2) our update model can generate guidance graphs for as large as $93 \\times 91$ maps and as many as 3,000 agents. We include the source code at: \\url{https://github.com/lunjohnzhang/ggo_public}. All optimized guidance graphs are available online at: \\url{https://yulunzhang.net/publication/zhang2024ggo}.", "date": "2024-02-02", "categories": []} {"id": "2404.16162", "title": "Scaling Lifelong Multi-Agent Path Finding to More Realistic Settings: Research Challenges and Opportunities", "abstract": "Multi-Agent Path Finding (MAPF) is the problem of moving multiple agents from starts to goals without collisions. Lifelong MAPF (LMAPF) extends MAPF by continuously assigning new goals to agents. We present our winning approach to the 2023 League of Robot Runners LMAPF competition, which leads us to several interesting research challenges and future directions. In this paper, we outline three main research challenges. The first challenge is to search for high-quality LMAPF solutions within a limited planning time (e.g., 1s per step) for a large number of agents (e.g., 10,000) or extremely high agent density (e.g., 97.7%). We present future directions such as developing more competitive rule-based and anytime MAPF algorithms and parallelizing state-of-the-art MAPF algorithms. The second challenge is to alleviate congestion and the effect of myopic behaviors in LMAPF algorithms. We present future directions, such as developing moving guidance and traffic rules to reduce congestion, incorporating future prediction and real-time search, and determining the optimal agent number. The third challenge is to bridge the gaps between the LMAPF models used in the literature and real-world applications. We present future directions, such as dealing with more realistic kinodynamic models, execution uncertainty, and evolving systems.", "date": "2024-04-24", "categories": []} {"id": "2510.03472", "title": "Destination-to-Chutes Task Mapping Optimization for Multi-Robot Coordination in Robotic Sorting Systems", "abstract": "We study optimizing a destination-to-chutes task mapping to improve throughput in Robotic Sorting Systems (RSS), where a team of robots sort packages on a sortation floor by transporting them from induct workstations to eject chutes based on their shipping destinations (e.g. Los Angeles or Pittsburgh). The destination-to-chutes task mapping is used to determine which chutes a robot can drop its package. Finding a high-quality task mapping is challenging because of the complexity of a real-world RSS. First, optimizing task mapping is interdependent with robot target assignment and path planning. Second, chutes will be CLOSED for a period of time once they receive sufficient packages to allow for downstream processing. Third, task mapping quality directly impacts the downstream processing, as scattered chutes for the same destination increase package handling time. In this paper, we first formally define task mappings and the problem of Task Mapping Optimization (TMO). We then present a simulator of RSS to evaluate task mappings. We then present a simple TMO method based on the Evolutionary Algorithm and Mixed Integer Linear Programming, demonstrating the advantage of our optimized task mappings over the greedily generated ones in various RSS setups with different map sizes, numbers of chutes, and destinations. Finally, we use Quality Diversity algorithms to analyze the throughput of a diverse set of task mappings. Our code is available online at https://github.com/lunjohnzhang/tmo_public.", "date": "2025-10-03", "categories": []} {"id": "2602.23579", "title": "Construct, Merge, Solve & Adapt with Reinforcement Learning for the min-max Multiple Traveling Salesman Problem", "abstract": "The Multiple Traveling Salesman Problem (mTSP) extends the Traveling Salesman Problem to m tours that start and end at a common depot and jointly visit all customers exactly once. In the min-max variant, the objective is to minimize the longest tour, reflecting workload balance. We propose a hybrid approach, Construct, Merge, Solve & Adapt with Reinforcement Learning (RL-CMSA), for the symmetric single-depot min-max mTSP. The method iteratively constructs diverse solutions using probabilistic clustering guided by learned pairwise q-values, merges routes into a compact pool, solves a restricted set-covering MILP, and refines solutions via inter-route remove, shift, and swap moves. The q-values are updated by reinforcing city-pair co-occurrences in high-quality solutions, while the pool is adapted through ageing and pruning. This combination of exact optimization and reinforcement-guided construction balances exploration and exploitation. Computational results on random and TSPLIB instances show that RL-CMSA consistently finds (near-)best solutions and outperforms a state-of-the-art hybrid genetic algorithm under comparable time limits, especially as instance size and the number of salesmen increase.", "date": "2026-02-27", "categories": []} {"id": "2602.24180", "title": "Learning Flexible Job Shop Scheduling under Limited Buffers and Material Kitting Constraints", "abstract": "The Flexible Job Shop Scheduling Problem (FJSP) originates from real production lines, while some practical constraints are often ignored or idealized in current FJSP studies, among which the limited buffer problem has a particular impact on production efficiency. To this end, we study an extended problem that is closer to practical scenarios--the Flexible Job Shop Scheduling Problem with Limited Buffers and Material Kitting. In recent years, deep reinforcement learning (DRL) has demonstrated considerable potential in scheduling tasks. However, its capacity for state modeling remains limited when handling complex dependencies and long-term constraints. To address this, we leverage a heterogeneous graph network within the DRL framework to model the global state. By constructing efficient message passing among machines, operations, and buffers, the network focuses on avoiding decisions that may cause frequent pallet changes during long-sequence scheduling, thereby helping improve buffer utilization and overall decision quality. Experimental results on both synthetic and real production line datasets show that the proposed method outperforms traditional heuristics and advanced DRL methods in terms of makespan and pallet changes, and also achieves a good balance between solution quality and computational cost. Furthermore, a supplementary video is provided to showcase a simulation system that effectively visualizes the progression of the production line.", "date": "2026-02-27", "categories": []} {"id": "2408.01503", "title": "Efficient Graph Coloring with Neural Networks: A Physics-Inspired Approach for Large Graphs", "abstract": "Combinatorial optimization problems near algorithmic phase transitions represent a fundamental challenge for both classical algorithms and machine learning approaches. Among them, graph coloring stands as a prototypical constraint satisfaction problem exhibiting sharp dynamical and satisfiability thresholds. Here we introduce a physics-inspired neural framework that learns to solve large-scale graph coloring instances by combining graph neural networks with statistical-mechanics principles. Our approach integrates a planting-based supervised signal, symmetry-breaking regularization, and iterative noise-annealed neural dynamics to navigate clustered solution landscapes. When the number of iterations scales quadratically with graph size, the learned solver reaches algorithmic thresholds close to the theoretical dynamical transition in random graphs and achieves near-optimal detection performance in the planted inference regime. The model generalizes from small training graphs to instances orders of magnitude larger, demonstrating that neural architectures can learn scalable algorithmic strategies that remain effective in hard connectivity regions. These results establish a general paradigm for learning neural solvers that operate near fundamental phase boundaries in combinatorial optimization and inference.", "date": "2024-08-02", "categories": []} {"id": "2602.21546", "title": "Mamba Meets Scheduling: Learning to Solve Flexible Job Shop Scheduling with Efficient Sequence Modeling", "abstract": "The Flexible Job Shop Problem (FJSP) is a well-studied combinatorial optimization problem with extensive applications for manufacturing and production scheduling. It involves assigning jobs to various machines to optimize criteria, such as minimizing total completion time. Current learning-based methods in this domain often rely on localized feature extraction models, limiting their capacity to capture overarching dependencies spanning operations and machines. This paper introduces an innovative architecture that harnesses Mamba, a state-space model with linear computational complexity, to facilitate comprehensive sequence modeling tailored for FJSP. In contrast to prevalent graph-attention-based frameworks that are computationally intensive for FJSP, we show our model is more efficient. Specifically, the proposed model possesses an encoder and a decoder. The encoder incorporates a dual Mamba block to extract operation and machine features separately. Additionally, we introduce an efficient cross-attention decoder to learn interactive embeddings of operations and machines. Our experimental results demonstrate that our method achieves faster solving speed and surpasses the performance of state-of-the-art learning-based methods for FJSP across various benchmarks.", "date": "2026-02-25", "categories": []} {"id": "2602.21995", "title": "Outpatient Appointment Scheduling Optimization with a Genetic Algorithm Approach", "abstract": "The optimization of complex medical appointment scheduling remains a significant operational challenge in multi-center healthcare environments, where clinical safety protocols and patient logistics must be reconciled. This study proposes and evaluates a Genetic Algorithm (GA) framework designed to automate the scheduling of multiple medical acts while adhering to rigorous inter-procedural incompatibility rules. Using a synthetic dataset encompassing 50 medical acts across four healthcare facilities, we compared two GA variants, Pre-Ordered and Unordered, against deterministic First-Come, First-Served (FCFS) and Random Choice baselines. Our results demonstrate that the GA framework achieved a 100% constraint fulfillment rate, effectively resolving temporal overlaps and clinical incompatibilities that the FCFS baseline failed to address in 60% and 40% of cases, respectively. Furthermore, the GA variants demonstrated statistically significant improvements (p < 0.001) in patient-centric metrics, achieving an Idle Time Ratio (ITR) frequently below 0.4 and reducing inter-healthcenter trips. While the GA (Ordered) variant provided a superior initial search locus, both evolutionary models converged to comparable global optima by the 100th generation. These findings suggest that transitioning from manual, human-mediated scheduling to an automated metaheuristic approach enhances clinical integrity, reduces administrative overhead, and significantly improves the patient experience by minimizing wait times and logistical burdens.", "date": "2026-02-25", "categories": []} {"id": "2602.18431", "title": "SMaRT: Online Reusable Resource Assignment and an Application to Mediation in the Kenyan Judiciary", "abstract": "Motivated by the problem of assigning mediators to cases in the Kenyan judicial, we study an online resource allocation problem where incoming tasks (cases) must be immediately assigned to available, capacity-constrained resources (mediators). The resources differ in their quality, which may need to be learned. In addition, resources can only be assigned to a subset of tasks that overlaps to varying degrees with the subset of tasks other resources can be assigned to. The objective is to maximize task completion while satisfying soft capacity constraints across all the resources. The scale of the real-world problem poses substantial challenges, since there are over 2000 mediators and a multitude of combinations of geographic locations (87) and case types (12) that each mediator is qualified to work on. Together, these features, unknown quality of new resources, soft capacity constraints, and a high-dimensional state space, make existing scheduling and resource allocation algorithms either inapplicable or inefficient. We formalize the problem in a tractable manner using a quadratic program formulation for assignment and a multi-agent bandit-style framework for learning. We demonstrate the key properties and advantages of our new algorithm, SMaRT (Selecting Mediators that are Right for the Task), compared with baselines on stylized instances of the mediator allocation problem. We then consider its application to real-world data on cases and mediators from the Kenyan judiciary. SMaRT outperforms baselines and allows control over the tradeoff between the strictness of capacity constraints and overall case resolution rates, both in settings where mediator quality is known beforehand and in bandit-like settings where learning is part of the problem definition. On the strength of these results, we plan to run a randomized controlled trial with SMaRT in the judiciary in the near future.", "date": "2026-02-20", "categories": []} {"id": "2602.20175", "title": "Tensor Network Generator-Enhanced Optimization for Traveling Salesman Problem", "abstract": "We present an application of the tensor network generator-enhanced optimization (TN-GEO) framework to address the traveling salesman problem (TSP), a fundamental combinatorial optimization challenge. Our approach employs a tensor network Born machine based on automatically differentiable matrix product states (MPS) as the generative model, using the Born rule to define probability distributions over candidate solutions. Unlike approaches based on binary encoding, which require $N^2$ variables and penalty terms to enforce valid tour constraints, we adopt a permutation-based formulation with integer variables and use autoregressive sampling with masking to guarantee that every generated sample is a valid tour by construction. We also introduce a $k$-site MPS variant that learns distributions over $k$-grams (consecutive city subsequences) using a sliding window approach, enabling parameter-efficient modeling for larger instances. Experimental validation on TSPLIB benchmark instances with up to 52 cities demonstrates that TN-GEO can outperform classical heuristics including swap and 2-opt hill-climbing. The $k$-site variants, which put more focus on local correlations, show better results compared to the full-MPS case.", "date": "2026-02-12", "categories": []} {"id": "2507.16874", "title": "Budget Allocation Policies for Real-Time Multi-Agent Path Finding", "abstract": "Multi-Agent Path finding (MAPF) is the problem of finding paths for a set of agents such that each agent reaches its desired destination while avoiding collisions with the other agents. This problem arises in many robotics applications, such as automated warehouses and swarms of drones. Many MAPF solvers are designed to run offline, that is, first generate paths for all agents and then execute them. In real-world scenarios, waiting for a complete solution before allowing any robot to move is often impractical. Real-time MAPF (RT-MAPF) captures this setting by assuming that agents must begin execution after a fixed planning period, referred to as the planning budget, and execute a fixed number of actions, referred to as the execution window. This results in an iterative process in which a short plan is executed, while the next execution window is planned concurrently. Existing solutions to RT-MAPF iteratively call windowed versions of MAPF algorithms in every planning period, without explicitly considering the size of the planning budget. We address this gap and explore different policies for allocating the planning budget in windowed versions of MAPF-LNS2, a state-of-the-art MAPF algorithm. Our exploration shows that the baseline approach in which all agents draw from a shared planning budget pool is ineffective in challenging scenarios. Instead, policies that intelligently distribute the planning budget among agents are able to solve more problem instances in less time.", "date": "2025-07-22", "categories": []} {"id": "2510.09469", "title": "Towards Information-Optimized Multi-Agent Path Finding: A Hybrid Framework with Reduced Inter-Agent Information Sharing", "abstract": "Multi-agent pathfinding (MAPF) remains a critical problem in robotics and autonomous systems, where agents must navigate shared spaces efficiently while avoiding conflicts. Traditional centralized algorithms with global information provide high-quality solutions but scale poorly in large-scale scenarios due to the combinatorial explosion of conflicts. Conversely, distributed approaches that have local information, particularly learning-based methods, offer better scalability by operating with relaxed information availability, yet often at the cost of solution quality. In realistic deployments, information is a constrained resource: broadcasting full agent states and goals can raise privacy concerns, strain limited bandwidth, and require extra sensing and communication hardware, increasing cost and energy use. We focus on the core question of how MAPF can be solved with minimal inter-agent information sharing while preserving solution feasibility. To this end, we present an information-centric formulation of the MAPF problem and introduce a hybrid framework, IO-MAPF, that integrates decentralized path planning with a lightweight centralized coordinator. In this framework, agents use reinforcement learning (RL) to plan independently, while the central coordinator provides minimal, targeted signals, such as static conflict-cell indicators or short conflict trajectories, that are dynamically shared to support efficient conflict resolution. We introduce an Information Units (IU) metric to quantify information use and show that our alert-driven design achieves 2x to 23x reduction in information sharing, compared to the state-of-the-art algorithms, while maintaining high success rates, demonstrating that reliable MAPF is achievable under strongly information-restricted, privacy-preserving conditions. We demonstrate the effectiveness of our algorithm using simulation and hardware experiments.", "date": "2025-10-10", "categories": []} {"id": "2506.22095", "title": "Beyond Simple Graphs: Neural Multi-Objective Routing on Multigraphs", "abstract": "Learning-based methods for routing have gained significant attention in recent years, both in single-objective and multi-objective contexts. Yet, existing methods are unsuitable for routing on multigraphs, which feature multiple edges with distinct attributes between node pairs, despite their strong relevance in real-world scenarios. In this paper, we propose two graph neural network-based methods to address multi-objective routing on multigraphs. Our first approach operates directly on the multigraph by autoregressively selecting edges until a tour is completed. The second model, which is more scalable, first simplifies the multigraph via a learned pruning strategy and then performs autoregressive routing on the resulting simple graph. We evaluate both models empirically, across a wide range of problems and graph distributions, and demonstrate their competitive performance compared to strong heuristics and neural baselines.", "date": "2025-06-27", "categories": []} {"id": "2503.12662", "title": "TuneNSearch: a hybrid transfer learning and local search approach for solving vehicle routing problems", "abstract": "This paper introduces TuneNSearch, a hybrid transfer learning and local search approach for addressing diverse variants of the vehicle routing problem (VRP). Our method uses reinforcement learning to generate high-quality solutions, which are subsequently refined by an efficient local search procedure. To ensure broad adaptability across VRP variants, TuneNSearch begins with a pre-training phase on the multi-depot VRP (MDVRP), followed by a fine-tuning phase to adapt it to other problem formulations. The learning phase utilizes a Transformer-based architecture enhanced with edge-aware attention, which integrates edge distances directly into the attention mechanism to better capture spatial relationships inherent to routing problems. We show that the pre-trained model generalizes effectively to single-depot variants, achieving performance comparable to models trained specifically on single-depot instances. Simultaneously, it maintains strong performance on multi-depot variants, an ability that models pre-trained solely on single-depot problems lack. For example, on 100-node instances of multi-depot variants, TuneNSearch outperforms a model pre-trained on the CVRP by 44%. In contrast, on 100-node instances of single-depot variants, TuneNSearch performs similar to the CVRP model. To validate the effectiveness of our method, we conduct extensive computational experiments on public benchmark and randomly generated instances. Across multiple CVRPLIB datasets, TuneNSearch consistently achieves performance deviations of less than 3% from the best-known solutions in the literature, compared to 6-25% for other neural-based models, depending on problem complexity. Overall, our approach demonstrates strong generalization to different problem sizes, instance distributions, and VRP formulations, while maintaining polynomial runtime complexity despite the integration of the local search algorithm.", "date": "2025-03-16", "categories": []} {"id": "2602.16012", "title": "Towards Efficient Constraint Handling in Neural Solvers for Routing Problems", "abstract": "Neural solvers have achieved impressive progress in addressing simple routing problems, particularly excelling in computational efficiency. However, their advantages under complex constraints remain nascent, for which current constraint-handling schemes via feasibility masking or implicit feasibility awareness can be inefficient or inapplicable for hard constraints. In this paper, we present Construct-and-Refine (CaR), the first general and efficient constraint-handling framework for neural routing solvers based on explicit learning-based feasibility refinement. Unlike prior construction-search hybrids that target reducing optimality gaps through heavy improvements yet still struggle with hard constraints, CaR achieves efficient constraint handling by designing a joint training framework that guides the construction module to generate diverse and high-quality solutions well-suited for a lightweight improvement process, e.g., 10 steps versus 5k steps in prior work. Moreover, CaR presents the first use of construction-improvement-shared representation, enabling potential knowledge sharing across paradigms by unifying the encoder, especially in more complex constrained scenarios. We evaluate CaR on typical hard routing constraints to showcase its broader applicability. Results demonstrate that CaR achieves superior feasibility, solution quality, and efficiency compared to both classical and neural state-of-the-art solvers.", "date": "2026-02-17", "categories": []} {"id": "2602.15070", "title": "An effective Genetic Programming Hyper-Heuristic for Uncertain Agile Satellite Scheduling", "abstract": "This paper investigates a novel problem, namely the Uncertain Agile Earth Observation Satellite Scheduling Problem (UAEOSSP). Unlike the static AEOSSP, it takes into account a range of uncertain factors (e.g., task profit, resource consumption, and task visibility) in order to reflect the reality that the actual information is inherently unknown beforehand. An effective Genetic Programming Hyper-Heuristic (GPHH) is designed to automate the generation of scheduling policies. The evolved scheduling policies can be utilized to adjust plans in real time and perform exceptionally well. Experimental results demonstrate that evolved scheduling policies significantly outperform both well-designed Look-Ahead Heuristics (LAHs) and Manually Designed Heuristics (MDHs). Specifically, the policies generated by GPHH achieve an average improvement of 5.03% compared to LAHs and 8.14% compared to MDHs.", "date": "2026-02-15", "categories": []} {"id": "2602.15721", "title": "Lifelong Scalable Multi-Agent Realistic Testbed and A Comprehensive Study on Design Choices in Lifelong AGV Fleet Management Systems", "abstract": "We present Lifelong Scalable Multi-Agent Realistic Testbed (LSMART), an open-source simulator to evaluate any Multi-Agent Path Finding (MAPF) algorithm in a Fleet Management System (FMS) with Automated Guided Vehicles (AGVs). MAPF aims to move a group of agents from their corresponding starting locations to their goals. Lifelong MAPF (LMAPF) is a variant of MAPF that continuously assigns new goals for agents to reach. LMAPF applications, such as autonomous warehouses, often require a centralized, lifelong system to coordinate the movement of a fleet of robots, typically AGVs. However, existing works on MAPF and LMAPF often assume simplified kinodynamic models, such as pebble motion, as well as perfect execution and communication for AGVs. Prior work has presented SMART, a software capable of evaluating any MAPF algorithms while considering agent kinodynamics, communication delays, and execution uncertainties. However, SMART is designed for MAPF, not LMAPF. Generalizing SMART to an FMS requires many more design choices. First, an FMS parallelizes planning and execution, raising the question of when to plan. Second, given planners with varying optimality and differing agent-model assumptions, one must decide how to plan. Third, when the planner fails to return valid solutions, the system must determine how to recover. In this paper, we first present LSMART, an open-source simulator that incorporates all these considerations to evaluate any MAPF algorithms in an FMS. We then provide experiment results based on state-of-the-art methods for each design choice, offering guidance on how to effectively design centralized lifelong AGV Fleet Management Systems. LSMART is available at https://smart-mapf.github.io/lifelong-smart.", "date": "2026-02-17", "categories": []} {"id": "2510.16701", "title": "An Agentic Framework with LLMs for Solving Complex Vehicle Routing Problems", "abstract": "Complex vehicle routing problems (VRPs) remain a fundamental challenge, demanding substantial expert effort for intent interpretation and algorithm design. While large language models (LLMs) offer a promising path toward automation, current approaches still rely on external intervention, which restrict autonomy and often lead to execution errors and low solution feasibility. To address these challenges, we propose an Agentic Framework with LLMs (AFL) for solving complex vehicle routing problems, achieving full automation from problem instance to solution. AFL directly extracts knowledge from raw inputs and enables self-contained code generation without handcrafted modules or external solvers. To improve trustworthiness, AFL decomposes the overall pipeline into three manageable subtasks and employs four specialized agents whose coordinated interactions enforce cross-functional consistency and logical soundness. Extensive experiments on 60 complex VRPs, ranging from standard benchmarks to practical variants, validate the effectiveness and generality of our framework, showing comparable performance against meticulously designed algorithms. Notably, it substantially outperforms existing LLM-based baselines in both code reliability and solution feasibility, achieving rates close to 100% on the evaluated benchmarks.", "date": "2025-10-19", "categories": []} {"id": "2602.05119", "title": "Unbiased Single-Queried Gradient for Combinatorial Objective", "abstract": "In a probabilistic reformulation of a combinatorial problem, we often face an optimization over a hypercube, which corresponds to the Bernoulli probability parameter for each binary variable in the primal problem. The combinatorial nature suggests that an exact gradient computation requires multiple queries. We propose a stochastic gradient that is unbiased and requires only a single query of the combinatorial function. This method encompasses a well-established REINFORCE (through an importance sampling), as well as including a class of new stochastic gradients.", "date": "2026-02-04", "categories": []} {"id": "2602.14676", "title": "GREAT-EER: Graph Edge Attention Network for Emergency Evacuation Responses", "abstract": "Emergency situations that require the evacuation of urban areas can arise from man-made causes (e.g., terrorist attacks or industrial accidents) or natural disasters, the latter becoming more frequent due to climate change. As a result, effective and fast methods to develop evacuation plans are of great importance. In this work, we identify and propose the Bus Evacuation Orienteering Problem (BEOP), an NP-hard combinatorial optimization problem with the goal of evacuating as many people from an affected area by bus in a short, predefined amount of time. The purpose of bus-based evacuation is to reduce congestion and disorder that arises in purely car-focused evacuation scenarios. To solve the BEOP, we propose a deep reinforcement learning-based method utilizing graph learning, which, once trained, achieves fast inference speed and is able to create evacuation routes in fractions of seconds. We can bound the gap of our evacuation plans using an MILP formulation. To validate our method, we create evacuation scenarios for San Francisco using real-world road networks and travel times. We show that we achieve near-optimal solution quality and are further able to investigate how many evacuation vehicles are necessary to achieve certain bus-based evacuation quotas given a predefined evacuation time while keeping run time adequate.", "date": "2026-02-16", "categories": []} {"id": "2602.14772", "title": "Learning Structural Hardness for Combinatorial Auctions: Instance-Dependent Algorithm Selection via Graph Neural Networks", "abstract": "The Winner Determination Problem (WDP) in combinatorial auctions is NP-hard, and no existing method reliably predicts which instances will defeat fast greedy heuristics. The ML-for-combinatorial-optimization community has focused on learning to \\emph{replace} solvers, yet recent evidence shows that graph neural networks (GNNs) rarely outperform well-tuned classical methods on standard benchmarks. We pursue a different objective: learning to predict \\emph{when} a given instance is hard for greedy allocation, enabling instance-dependent algorithm selection. We design a 20-dimensional structural feature vector and train a lightweight MLP hardness classifier that predicts the greedy optimality gap with mean absolute error 0.033, Pearson correlation 0.937, and binary classification accuracy 94.7\\% across three random seeds. For instances identified as hard -- those exhibiting ``whale-fish'' trap structure where greedy provably fails -- we deploy a heterogeneous GNN specialist that achieves ${\\approx}0\\%$ optimality gap on all six adversarial configurations tested (vs.\\ 3.75--59.24\\% for greedy). A hybrid allocator combining the hardness classifier with GNN and greedy solvers achieves 0.51\\% overall gap on mixed distributions. Our honest evaluation on CATS benchmarks confirms that GNNs do not outperform Gurobi (0.45--0.71 vs.\\ 0.20 gap), motivating the algorithm selection framing. Learning \\emph{when} to deploy expensive solvers is more tractable than learning to replace them.", "date": "2026-02-16", "categories": []} {"id": "2508.11850", "title": "EvoCut: Strengthening Integer Programs via Evolution-Guided Language Models", "abstract": "Integer programming (IP) is central to many combinatorial optimization tasks but remains challenging due to its NP-hard nature. A practical way to improve IP solvers is to manually design acceleration cuts, i.e., inequalities that speed up solving. However, this creative process requires deep expertise and has been difficult to automate. Our proposed framework, EvoCut, automates the generation of acceleration cuts at the symbolic modeling level: it reasons over a symbolic MILP model and a natural language description of the problem to discover a reusable set of acceleration cuts that can be used for each concrete instance of the model. EvoCut (i) initializes a population of candidate cuts via an initializer agent that uses an LLM, (ii) empirically screens candidates on a small verification set by checking that reference solutions remain feasible and that at least one stored LP relaxation solution is cut off, and (iii) iteratively refines the population through evolutionary crossover and mutation agents. Compared to baseline MILP formulations solved with a fixed time budget, EvoCut reduces optimality gaps by up to $76\\%$ and reaches target gaps up to $7.2$ times faster (shifted geometric mean speedup). Ablations show its robustness across different LLM backends and across solvers/cut settings. Code: https://github.com/milad1378yz/EvoCut.", "date": "2025-08-16", "categories": []} {"id": "2602.11346", "title": "Divide and Learn: Multi-Objective Combinatorial Optimization at Scale", "abstract": "Multi-objective combinatorial optimization seeks Pareto-optimal solutions over exponentially large discrete spaces, yet existing methods sacrifice generality, scalability, or theoretical guarantees. We reformulate it as an online learning problem over a decomposed decision space, solving position-wise bandit subproblems via adaptive expert-guided sequential construction. This formulation admits regret bounds of $O(d\\sqrt{T \\log T})$ depending on subproblem dimensionality \\(d\\) rather than combinatorial space size. On standard benchmarks, our method achieves 80--98\\% of specialized solvers performance while achieving two to three orders of magnitude improvement in sample and computational efficiency over Bayesian optimization methods. On real-world hardware-software co-design for AI accelerators with expensive simulations, we outperform competing methods under fixed evaluation budgets. The advantage grows with problem scale and objective count, establishing bandit optimization over decomposed decision spaces as a principled alternative to surrogate modeling or offline training for multi-objective optimization.", "date": "2026-02-11", "categories": []} {"id": "2602.12055", "title": "Multi UAVs Preflight Planning in a Shared and Dynamic Airspace", "abstract": "Preflight planning for large-scale Unmanned Aerial Vehicle (UAV) fleets in dynamic, shared airspace presents significant challenges, including temporal No-Fly Zones (NFZs), heterogeneous vehicle profiles, and strict delivery deadlines. While Multi-Agent Path Finding (MAPF) provides a formal framework, existing methods often lack the scalability and flexibility required for real-world Unmanned Traffic Management (UTM). We propose DTAPP-IICR: a Delivery-Time Aware Prioritized Planning method with Incremental and Iterative Conflict Resolution. Our framework first generates an initial solution by prioritizing missions based on urgency. Secondly, it computes roundtrip trajectories using SFIPP-ST, a novel 4D single-agent planner (Safe Flight Interval Path Planning with Soft and Temporal Constraints). SFIPP-ST handles heterogeneous UAVs, strictly enforces temporal NFZs, and models inter-agent conflicts as soft constraints. Subsequently, an iterative Large Neighborhood Search, guided by a geometric conflict graph, efficiently resolves any residual conflicts. A completeness-preserving directional pruning technique further accelerates the 3D search. On benchmarks with temporal NFZs, DTAPP-IICR achieves near-100% success with fleets of up to 1,000 UAVs and gains up to 50% runtime reduction from pruning, outperforming batch Enhanced Conflict-Based Search in the UTM context. Scaling successfully in realistic city-scale operations where other priority-based methods fail even at moderate deployments, DTAPP-IICR is positioned as a practical and scalable solution for preflight planning in dense, dynamic urban airspace.", "date": "2026-02-12", "categories": []} {"id": "2602.10794", "title": "Transport, Don't Generate: Deterministic Geometric Flows for Combinatorial Optimization", "abstract": "Recent advances in Neural Combinatorial Optimization (NCO) have been dominated by diffusion models that treat the Euclidean Traveling Salesman Problem (TSP) as a stochastic $N \\times N$ heatmap generation task. In this paper, we propose CycFlow, a framework that replaces iterative edge denoising with deterministic point transport. CycFlow learns an instance-conditioned vector field that continuously transports input 2D coordinates to a canonical circular arrangement, where the optimal tour is recovered from this $2N$ dimensional representation via angular sorting. By leveraging data-dependent flow matching, we bypass the quadratic bottleneck of edge scoring in favor of linear coordinate dynamics. This paradigm shift accelerates solving speed by up to three orders of magnitude compared to state-of-the-art diffusion baselines, while maintaining competitive optimality gaps.", "date": "2026-02-11", "categories": []} {"id": "2602.11057", "title": "Divide, Harmonize, Then Conquer It: Shooting Multi-Commodity Flow Problems with Multimodal Language Models", "abstract": "The multi-commodity flow (MCF) problem is a fundamental topic in network flow and combinatorial optimization, with broad applications in transportation, communication, and logistics, etc. Nowadays, the rapid expansion of allocation systems has posed challenges for existing optimization engines in balancing optimality and tractability. In this paper, we present Pram, the first ML-based method that leverages the reasoning power of multimodal language models (MLMs) for addressing the trade-off dilemma -- a great need of service providers. As part of our proposal, Pram (i) quickly computes high-quality allocations by dividing the original problem into local subproblems, which are then resolved by an MLM-powered \"agent\", and (ii) ensures global consistency by harmonizing these subproblems via a multi-agent reinforcement learning algorithm. Theoretically, we show that Pram, which learns to perform gradient descent in context, provably converges to the optimum within the family of MCF problems. Empirically, on real-world datasets and public topologies, Pram achieves performance comparable to, and in some cases even surpassing, linear programming solvers (very close to the optimal solution), and substantially lower runtimes (1 to 2 orders of magnitude faster). Moreover, Pram exhibits strong robustness (<10\\% performance degradation under link failures or flow bursts), demonstrating MLM's generalization ability to unforeseen events. Pram is objective-agnostic and seamlessly integrates with mainstream allocation systems, providing a practical and scalable solution for future networks.", "date": "2026-02-11", "categories": []} {"id": "2602.10037", "title": "Effectiveness of Binary Autoencoders for QUBO-Based Optimization Problems", "abstract": "In black-box combinatorial optimization, objective evaluations are often expensive, so high quality solutions must be found under a limited budget. Factorization machine with quantum annealing (FMQA) builds a quadratic surrogate model from evaluated samples and optimizes it on an Ising machine. However, FMQA requires binary decision variables, and for nonbinary structures such as integer permutations, the choice of binary encoding strongly affects search efficiency. If the encoding fails to reflect the original neighborhood structure, small Hamming moves may not correspond to meaningful modifications in the original solution space, and constrained problems can yield many infeasible candidates that waste evaluations. Recent work combines FMQA with a binary autoencoder (bAE) that learns a compact binary latent code from feasible solutions, yet the mechanism behind its performance gains is unclear. Using a small traveling salesman problem as an interpretable testbed, we show that the bAE reconstructs feasible tours accurately and, compared with manually designed encodings at similar compression, better aligns tour distances with latent Hamming distances, yields smoother neighborhoods under small bit flips, and produces fewer local optima. These geometric properties explain why bAE+FMQA improves the approximation ratio faster while maintaining feasibility throughout optimization, and they provide guidance for designing latent representations for black-box optimization.", "date": "2026-02-10", "categories": []} {"id": "2509.21765", "title": "Lifelong Learning with Behavior Consolidation for Vehicle Routing", "abstract": "Recent neural solvers have demonstrated promising performance in learning to solve routing problems. However, existing studies are primarily based on one-off training on one or a set of predefined problem distributions and scales, i.e., tasks. When a new task arises, they typically rely on either zero-shot generalization, which may be poor due to the discrepancies between the new task and the training task(s), or fine-tuning the pretrained solver on the new task, which possibly leads to catastrophic forgetting of knowledge acquired from previous tasks. This paper explores a novel lifelong learning paradigm for neural VRP solvers, where multiple tasks with diverse distributions and scales arise sequentially over time. Solvers are required to effectively and efficiently learn to solve new tasks while maintaining their performance on previously learned tasks. Consequently, a novel framework called Lifelong Learning Router with Behavior Consolidation (LLR-BC) is proposed. LLR-BC consolidates prior knowledge effectively by aligning behaviors of the solver trained on a new task with the buffered ones in a decision-seeking way. To encourage more focus on crucial experiences, LLR-BC assigns greater consolidated weights to decisions with lower confidence. Extensive experiments on capacitated vehicle routing problems and traveling salesman problems demonstrate LLR-BC's effectiveness in training high-performance neural solvers in a lifelong learning setting, addressing the catastrophic forgetting issue, maintaining their plasticity, and improving zero-shot generalization ability.", "date": "2025-09-26", "categories": []} {"id": "2601.22896", "title": "Game-Theoretic Co-Evolution for LLM-Based Heuristic Discovery", "abstract": "Large language models (LLMs) have enabled rapid progress in automatic heuristic discovery (AHD), yet most existing methods are predominantly limited by static evaluation against fixed instance distributions, leading to potential overfitting and poor generalization under distributional shifts. We propose Algorithm Space Response Oracles (ASRO), a game-theoretic framework that reframes heuristic discovery as a program level co-evolution between solver and instance generator. ASRO models their interaction as a two-player zero-sum game, maintains growing strategy pools on both sides, and iteratively expands them via LLM-based best-response oracles against mixed opponent meta-strategies, thereby replacing static evaluation with an adaptive, self-generated curriculum. Across multiple combinatorial optimization domains, ASRO consistently outperforms static-training AHD baselines built on the same program search mechanisms, achieving substantially improved generalization and robustness on diverse and out-of-distribution instances.", "date": "2026-01-30", "categories": []} {"id": "2602.07216", "title": "Probing Neural TSP Representations for Prescriptive Decision Support", "abstract": "The field of neural combinatorial optimization (NCO) trains neural policies to solve NP-hard problems such as the traveling salesperson problem (TSP). We ask whether, beyond producing good tours, a trained TSP solver learns internal representations that transfer to other optimization-relevant objectives, in the spirit of transfer learning from other domains. We train several attention-based TSP policies, collect their internal activations, and train probes on node/edge embeddings for two NP-hard prescriptive downstream tasks inspired by real-world logistics scenarios: node-removal sensitivity (identifying the most impactful node to remove) and edge-forbid sensitivity (identifying the most critical edge to retain). On a Euclidean TSP100-trained model, probes for both tasks are competitive with existing baselines. Ensembling probe signals with geometric features outperforms the strongest baselines: 65\\% top-1 accuracy (vs. 58\\% baseline) for the best-node-removal task, and 73\\% top-1 accuracy (vs. 67\\% baseline) for the worst-edge identification task. To our knowledge, we are the first to study neural TSP solvers as transferable encoders for prescriptive what-if decision-support objectives beyond tour construction. Finally, we show that transfer accuracy increases with solver quality across training and model scale, suggesting that training stronger NCO solvers also yields more useful encoders for downstream objectives. Our code is available at: github.com/ReubenNarad/tsp_prescriptive_probe", "date": "2026-02-06", "categories": []} {"id": "2602.08052", "title": "Graph-Enhanced Deep Reinforcement Learning for Multi-Objective Unrelated Parallel Machine Scheduling", "abstract": "The Unrelated Parallel Machine Scheduling Problem (UPMSP) with release dates, setups, and eligibility constraints presents a significant multi-objective challenge. Traditional methods struggle to balance minimizing Total Weighted Tardiness (TWT) and Total Setup Time (TST). This paper proposes a Deep Reinforcement Learning framework using Proximal Policy Optimization (PPO) and a Graph Neural Network (GNN). The GNN effectively represents the complex state of jobs, machines, and setups, allowing the PPO agent to learn a direct scheduling policy. Guided by a multi-objective reward function, the agent simultaneously minimizes TWT and TST. Experimental results on benchmark instances demonstrate that our PPO-GNN agent significantly outperforms a standard dispatching rule and a metaheuristic, achieving a superior trade-off between both objectives. This provides a robust and scalable solution for complex manufacturing scheduling.", "date": "2026-02-08", "categories": []} {"id": "2602.08253", "title": "G-LNS: Generative Large Neighborhood Search for LLM-Based Automatic Heuristic Design", "abstract": "While Large Language Models (LLMs) have recently shown promise in Automated Heuristic Design (AHD), existing approaches typically formulate AHD around constructive priority rules or parameterized local search guidance, thereby restricting the search space to fixed heuristic forms. Such designs offer limited capacity for structural exploration, making it difficult to escape deep local optima in complex Combinatorial Optimization Problems (COPs). In this work, we propose G-LNS, a generative evolutionary framework that extends LLM-based AHD to the automated design of Large Neighborhood Search (LNS) operators. Unlike prior methods that evolve heuristics in isolation, G-LNS leverages LLMs to co-evolve tightly coupled pairs of destroy and repair operators. A cooperative evaluation mechanism explicitly captures their interaction, enabling the discovery of complementary operator logic that jointly performs effective structural disruption and reconstruction. Extensive experiments on challenging COP benchmarks, such as Traveling Salesman Problems (TSP) and Capacitated Vehicle Routing Problems (CVRP), demonstrate that G-LNS significantly outperforms LLM-based AHD methods as well as strong classical solvers. The discovered heuristics not only achieve near-optimal solutions with reduced computational budgets but also exhibit robust generalization across diverse and unseen instance distributions.", "date": "2026-02-09", "categories": []} {"id": "2602.08297", "title": "Automatic Generation of Polynomial Symmetry Breaking Constraints", "abstract": "Symmetry in integer programming causes redundant search and is often handled with symmetry breaking constraints that remove as many equivalent solutions as possible. We propose an algebraic method which allows to generate a random family of polynomial inequalities which can be used as symmetry breakers. The method requires as input an arbitrary base polynomial and a group of permutations which is specific to the integer program. The computations can be easily carried out in any major symbolic computation software. In order to test our approach, we describe a case study on near half-capacity 0-1 bin packing instances which exhibit substantial symmetries. We statically generate random quadratic breakers and add them to a baseline integer programming problem which we then solve with Gurobi. It turns out that simple symmetry breakers, especially combining few variables and permutations, most consistently reduce work time.", "date": "2026-02-09", "categories": []} {"id": "2602.04989", "title": "Near-Optimal Dynamic Matching via Coarsening with Application to Heart Transplantation", "abstract": "Online matching has been a mainstay in domains such as Internet advertising and organ allocation, but practical algorithms often lack strong theoretical guarantees. We take an important step toward addressing this by developing new online matching algorithms based on a coarsening approach. Although coarsening typically implies a loss of granularity, we show that, to the contrary, aggregating offline nodes into capacitated clusters can yield near-optimal theoretical guarantees. We apply our methodology to heart transplant allocation to develop theoretically grounded policies based on structural properties of historical data. Furthermore, in simulations based on real data, our policy closely matches the performance of the omniscient benchmark, achieving competitive ratio 0.91, drastically higher than the US status quo policy's 0.51. Our work bridges the gap between data-driven heuristics and pessimistic theoretical lower bounds.", "date": "2026-02-04", "categories": []} {"id": "2602.06064", "title": "iScheduler: Reinforcement Learning-Driven Continual Optimization for Large-Scale Resource Investment Problems", "abstract": "Scheduling precedence-constrained tasks under shared renewable resources is central to modern computing platforms. The Resource Investment Problem (RIP) models this setting by minimizing the cost of provisioned renewable resources under precedence and timing constraints. Exact mixed-integer programming and constraint programming become impractically slow on large instances, and dynamic updates require schedule revisions under tight latency budgets. We present iScheduler, a reinforcement-learning-driven iterative scheduling framework that formulates RIP solving as a Markov decision process over decomposed subproblems and constructs schedules through sequential process selection. The framework accelerates optimization and supports reconfiguration by reusing unchanged process schedules and rescheduling only affected processes. We also release L-RIPLIB, an industrial-scale benchmark derived from cloud-platform workloads with 1,000 instances of 2,500-10,000 tasks. Experiments show that iScheduler attains competitive resource costs while reducing time to feasibility by up to 43$\\times$ against strong commercial baselines.", "date": "2026-01-30", "categories": []} {"id": "2602.06693", "title": "Makespan Minimization in Split Learning: From Theory to Practice", "abstract": "Split learning recently emerged as a solution for distributed machine learning with heterogeneous IoT devices, where clients can offload part of their training to computationally-powerful helpers. The core challenge in split learning is to minimize the training time by jointly devising the client-helper assignment and the schedule of tasks at the helpers. We first study the model where each helper has a memory cardinality constraint on how many clients it may be assigned, which represents the case of homogeneous tasks. Through complexity theory, we rule out exact polynomial-time algorithms and approximation schemes even for highly restricted instances of this problem. We complement these negative results with a non-trivial polynomial-time 5-approximation algorithm. Building on this, we then focus on the more general heterogeneous task setting considered by Tirana et al. [INFOCOM 2024], where helpers have memory capacity constraints and clients have variable memory costs. In this case, we prove that, unless P=NP, the problem cannot admit a polynomial-time approximation algorithm for any approximation factor. However, by adapting our aforementioned 5-approximation algorithm, we develop a novel heuristic for the heterogeneous task setting and show that it outperforms heuristics from prior works through extensive experiments.", "date": "2026-02-06", "categories": []} {"id": "2505.24779", "title": "Are Your Generated Instances Truly Useful? GenBench-MILP: A Benchmark Suite for MILP Instance Generation", "abstract": "The proliferation of machine learning-based methods for Mixed-Integer Linear Programming (MILP) instance generation has surged, driven by the need for diverse training datasets. However, a critical question remains: Are these generated instances truly useful and realistic? Current evaluation protocols often rely on superficial structural metrics or simple solvability checks, which frequently fail to capture the true computational complexity of real-world problems. To bridge this gap, we introduce GenBench-MILP, a comprehensive benchmark suite designed for the standardized and objective evaluation of MILP generators. Our framework assesses instance quality across four key dimensions: mathematical validity, structural similarity, computational hardness, and utility in downstream tasks. A distinctive innovation of GenBench-MILP is the analysis of solver-internal features -- including root node gaps, heuristic success rates, and cut plane usage. By treating the solver's dynamic behavior as an expert assessment, we reveal nuanced computational discrepancies that static graph features miss. Our experiments on instance generative models demonstrate that instances with high structural similarity scores can still exhibit drastically divergent solver interactions and difficulty levels. By providing this multifaceted evaluation toolkit, GenBench-MILP aims to facilitate rigorous comparisons and guide the development of high-fidelity instance generators.", "date": "2025-05-30", "categories": []} {"id": "2602.05875", "title": "Beyond Manual Planning: Seating Allocation for Large Organizations", "abstract": "We introduce the Hierarchical Seating Allocation Problem (HSAP) which addresses the optimal assignment of hierarchically structured organizational teams to physical seating arrangements on a floor plan. This problem is driven by the necessity for large organizations with large hierarchies to ensure that teams with close hierarchical relationships are seated in proximity to one another, such as ensuring a research group occupies a contiguous area. Currently, this problem is managed manually leading to infrequent and suboptimal replanning efforts. To alleviate this manual process, we propose an end-to-end framework to solve the HSAP. A scalable approach to calculate the distance between any pair of seats using a probabilistic road map (PRM) and rapidly-exploring random trees (RRT) which is combined with heuristic search and dynamic programming approach to solve the HSAP using integer programming. We demonstrate our approach under different sized instances by evaluating the PRM framework and subsequent allocations both quantitatively and qualitatively.", "date": "2026-02-05", "categories": []} {"id": "2508.18742", "title": "Constraint Matters: Multi-Modal Representation for Reducing Mixed-Integer Linear programming", "abstract": "Model reduction, which aims to learn a simpler model of the original mixed integer linear programming (MILP), can solve large-scale MILP problems much faster. Most existing model reduction methods are based on variable reduction, which predicts a solution value for a subset of variables. From a dual perspective, constraint reduction that transforms a subset of inequality constraints into equalities can also reduce the complexity of MILP, but has been largely ignored. Therefore, this paper proposes a novel constraint-based model reduction approach for the MILP. Constraint-based MILP reduction has two challenges: 1) which inequality constraints are critical such that reducing them can accelerate MILP solving while preserving feasibility, and 2) how to predict these critical constraints efficiently. To identify critical constraints, we first label these tight-constraints at the optimal solution as potential critical constraints and design a heuristic rule to select a subset of critical tight-constraints. To learn the critical tight-constraints, we propose a multi-modal representation technique that leverages information from both instance-level and abstract-level MILP formulations. The experimental results show that, compared to the state-of-the-art methods, our method improves the quality of the solution by over 50\\% and reduces the computation time by 17.47\\%.", "date": "2025-08-26", "categories": []} {"id": "2601.19395", "title": "SEAFormer: A Spatial Proximity and Edge-Aware Transformer for Real-World Vehicle Routing Problems", "abstract": "Real-world Vehicle Routing Problems (RWVRPs) require solving complex, sequence-dependent challenges at scale with constraints such as delivery time window, replenishment or recharging stops, asymmetric travel cost, etc. While recent neural methods achieve strong results on large-scale classical VRP benchmarks, they struggle to address RWVRPs because their strategies overlook sequence dependencies and underutilize edge-level information, which are precisely the characteristics that define the complexity of RWVRPs. We present SEAFormer, a novel transformer that incorporates both node-level and edge-level information in decision-making through two key innovations. First, our Clustered Proximity Attention (CPA) exploits locality-aware clustering to reduce the complexity of attention from $O(n^2)$ to $O(n)$ while preserving global perspective, allowing SEAFormer to efficiently train on large instances. Second, our lightweight edge-aware module captures pairwise features through residual fusion, enabling effective incorporation of edge-based information and faster convergence. Extensive experiments across four RWVRP variants with various scales demonstrate that SEAFormer achieves superior results over state-of-the-art methods. Notably, SEAFormer is the first neural method to solve 1,000+ node RWVRPs effectively, while also achieving superior performance on classic VRPs, making it a versatile solution for both research benchmarks and real-world applications.", "date": "2026-01-27", "categories": []} {"id": "2602.00408", "title": "Variational Approach for Job Shop Scheduling", "abstract": "This paper proposes a novel Variational Graph-to-Scheduler (VG2S) framework for solving the Job Shop Scheduling Problem (JSSP), a critical task in manufacturing that directly impacts operational efficiency and resource utilization. Conventional Deep Reinforcement Learning (DRL) approaches often face challenges such as non-stationarity during training and limited generalization to unseen problem instances because they optimize representation learning and policy execution simultaneously. To address these issues, we introduce variational inference to the JSSP domain for the first time and derive a probabilistic objective based on the Evidence of Lower Bound (ELBO) with maximum entropy reinforcement learning. By mathematically decoupling representation learning from policy optimization, the VG2S framework enables the agent to learn robust structural representations of scheduling instances through a variational graph encoder. This approach significantly enhances training stability and robustness against hyperparameter variations. Extensive experiments demonstrate that the proposed method exhibits superior zero-shot generalization compared with state-of-the-art DRL baselines and traditional dispatching rules, particularly on large-scale and challenging benchmark instances such as DMU and SWV.", "date": "2026-01-30", "categories": []} {"id": "2602.00488", "title": "OD-DEAL: Dynamic Expert-Guided Adversarial Learning with Online Decomposition for Scalable Capacitated Vehicle Routing", "abstract": "Solving large-scale capacitated vehicle routing problems (CVRP) is hindered by the high complexity of heuristics and the limited generalization of neural solvers on massive graphs. We propose OD-DEAL, an adversarial learning framework that tightly integrates hybrid genetic search (HGS) and online barycenter clustering (BCC) decomposition, and leverages high-fidelity knowledge distillation to transfer expert heuristic behavior. OD-DEAL trains a graph attention network (GAT)-based generative policy through a minimax game, in which divide-and-conquer strategies from a hybrid expert are distilled into dense surrogate rewards. This enables high-quality, clustering-free inference on large-scale instances. Empirical results demonstrate that OD-DEAL achieves state-of-the-art (SOTA) real-time CVRP performance, solving 10000-node instances with near-constant neural scaling. This uniquely enables the sub-second, heuristic-quality inference required for dynamic large-scale deployment.", "date": "2026-01-31", "categories": []} {"id": "2602.02736", "title": "Time-Critical Multimodal Medical Transportation: Organs, Patients, and Medical Supplies", "abstract": "Timely transportation of organs, patients, and medical supplies is critical to modern healthcare, particularly in emergencies and transplant scenarios where even short delays can severely impact outcomes. Traditional ground-based vehicles such as ambulances are often hindered by traffic congestion; while air vehicles such as helicopters are faster but costly. Emerging air vehicles -- Unmanned Aerial Vehicles and electric vertical take-off and landing aircraft -- have lower operating costs, but remain limited by range and susceptibility to weather conditions. A multimodal transportation system that integrates both air and ground vehicles can leverage the strengths of each to enhance overall transportation efficiency. This study introduces a constructive greedy heuristic algorithm for multimodal vehicle dispatching for medical transportation. Four different fleet configurations were tested: (i) ambulances only, (ii) ambulances with Unmanned Aerial Vehicles, (iii) ambulances with electric vertical take-off and landing aircraft, and (iv) a fully integrated fleet of ambulances, Unmanned Aerial Vehicles, and electric vertical take-off and landing aircraft. The algorithm incorporates payload consolidation across compatible routes, accounts for traffic congestion in ground operations and weather conditions in aerial operations, while enabling rapid vehicle dispatching compared to computationally intensive optimization models. Using a common set of conditions, we evaluate all four fleet types to identify the most effective configurations for fulfilling medical transportation needs while minimizing operating costs, recharging/fuel costs, and total transportation time.", "date": "2026-02-02", "categories": []} {"id": "2602.03588", "title": "Efficient Algorithms for Partial Constraint Satisfaction Problems over Control-flow Graphs", "abstract": "In this work, we focus on the Partial Constraint Satisfaction Problem (PCSP) over control-flow graphs (CFGs) of programs. PCSP serves as a generalization of the well-known Constraint Satisfaction Problem (CSP). In the CSP framework, we define a set of variables, a set of constraints, and a finite domain $D$ that encompasses all possible values for each variable. The objective is to assign a value to each variable in such a way that all constraints are satisfied. In the graph variant of CSP, an underlying graph is considered and we have one variable corresponding to each vertex of the graph and one or several constraints corresponding to each edge. In PCSPs, we allow for certain constraints to be violated at a specified cost, aiming to find a solution that minimizes the total cost. Numerous classical compiler optimization tasks can be framed as PCSPs over control-flow graphs. Examples include Register Allocation, Lifetime-optimal Speculative Partial Redundancy Elimination (LOSPRE), and Optimal Placement of Bank Selection Instructions. On the other hand, it is well-known that control-flow graphs of structured programs are sparse and decomposable in a variety of ways. In this work, we rely on the Series-Parallel-Loop (SPL) decompositions as introduced by~\\cite{RegisterAllocation}. Our main contribution is a general algorithm for PCSPs over SPL graphs with a time complexity of \\(O(|G| \\cdot |D|^6)\\), where \\(|G|\\) represents the size of the control-flow graph. Note that for any fixed domain $D,$ this yields a linear-time solution. Our algorithm can be seen as a generalization and unification of previous SPL-based approaches for register allocation and LOSPRE. In addition, we provide experimental results over another classical PCSP task, i.e. Optimal Bank Selection, achieving runtimes four times better than the previous state of the art.", "date": "2026-02-03", "categories": []} {"id": "2601.02649", "title": "Effective Online 3D Bin Packing with Lookahead Parcels Using Monte Carlo Tree Search", "abstract": "Online 3D Bin Packing (3D-BP) with robotic arms is crucial for reducing transportation and labor costs in modern logistics. While Deep Reinforcement Learning (DRL) has shown strong performance, it often fails to adapt to real-world short-term distribution shifts, which arise as different batches of goods arrive sequentially, causing performance drops. We argue that the short-term lookahead information available in modern logistics systems is key to mitigating this issue, especially during distribution shifts. We formulate online 3D-BP with lookahead parcels as a Model Predictive Control (MPC) problem and adapt the Monte Carlo Tree Search (MCTS) framework to solve it. Our framework employs a dynamic exploration prior that automatically balances a learned RL policy and a robust random policy based on the lookahead characteristics. Additionally, we design an auxiliary reward to penalize long-term spatial waste from individual placements. Extensive experiments on real-world datasets show that our method consistently outperforms state-of-the-art baselines, achieving over 10\\% gains under distributional shifts, 4\\% average improvement in online deployment, and up to more than 8\\% in the best case--demonstrating the effectiveness of our framework.", "date": "2026-01-06", "categories": []} {"id": "2601.13465", "title": "Graph Neural Networks are Heuristics", "abstract": "We demonstrate that a single training trajectory can transform a graph neural network into an unsupervised heuristic for combinatorial optimization. Focusing on the Travelling Salesman Problem, we show that encoding global structural constraints as an inductive bias enables a non-autoregressive model to generate solutions via direct forward passes, without search, supervision, or sequential decision-making. At inference time, dropout and snapshot ensembling allow a single model to act as an implicit ensemble, reducing optimality gaps through increased solution diversity. Our results establish that graph neural networks do not require supervised training nor explicit search to be effective. Instead, they can internalize global combinatorial structure and function as strong, learned heuristics. This reframes the role of learning in combinatorial optimization: from augmenting classical algorithms to directly instantiating new heuristics.", "date": "2026-01-19", "categories": []} {"id": "2602.00302", "title": "Neural Ising Machines via Unrolling and Zeroth-Order Training", "abstract": "We propose a data-driven heuristic for NP-hard Ising and Max-Cut optimization that learns the update rule of an iterative dynamical system. The method learns a shared, node-wise update rule that maps local interaction fields to spin updates, parameterized by a compact multilayer perceptron with a small number of parameters. Training is performed using a zeroth-order optimizer, since backpropagation through long, recurrent Ising-machine dynamics leads to unstable and poorly informative gradients. We call this approach a neural network parameterized Ising machine (NPIM). Despite its low parameter count, the learned dynamics recover effective algorithmic structure, including momentum-like behavior and time-varying schedules, enabling efficient search in highly non-convex energy landscapes. Across standard Ising and neural combinatorial optimization benchmarks, NPIM achieves competitive solution quality and time-to-solution relative to recent learning-based methods and strong classical Ising-machine heuristics.", "date": "2026-01-30", "categories": []} {"id": "2602.00549", "title": "Beyond the Node: Clade-level Selection for Efficient MCTS in Automatic Heuristic Design", "abstract": "While Monte Carlo Tree Search (MCTS) shows promise in Large Language Model (LLM) based Automatic Heuristic Design (AHD), it suffers from a critical over-exploitation tendency under the limited computational budgets required for heuristic evaluation. To address this limitation, we propose Clade-AHD, an efficient framework that replaces node-level point estimates with clade-level Bayesian beliefs. By aggregating descendant evaluations into Beta distributions and performing Thompson Sampling over these beliefs, Clade-AHD explicitly models uncertainty to guide exploration, enabling more reliable decision-making under sparse and noisy evaluations. Extensive experiments on complex combinatorial optimization problems demonstrate that Clade-AHD consistently outperforms state-of-the-art methods while significantly reducing computational cost. The source code is publicly available at: https://github.com/Mriya0306/Clade-AHD.", "date": "2026-01-31", "categories": []} {"id": "2602.00580", "title": "Small Shifts, Large Gains: Unlocking Traditional TSP Heuristic Guided-Sampling via Unsupervised Neural Instance Modification", "abstract": "The Traveling Salesman Problem (TSP) is one of the most representative NP-hard problems in route planning and a long-standing benchmark in combinatorial optimization. Traditional heuristic tour constructors, such as Farthest or Nearest Insertion, are computationally efficient and highly practical, but their deterministic behavior limits exploration and often leads to local optima. In contrast, neural-based heuristic tour constructors alleviate this issue through guided-sampling and typically achieve superior solution quality, but at the cost of extensive training and reliance on ground-truth supervision, hindering their practical use. To bridge this gap, we propose TSP-MDF, a novel instance modification framework that equips traditional deterministic heuristic tour constructors with guided-sampling capability. Specifically, TSP-MDF introduces a neural-based instance modifier that strategically shifts node coordinates to sample multiple modified instances, on which the base traditional heuristic tour constructor constructs tours that are mapped back to the original instance, allowing traditional tour constructors to explore higher-quality tours and escape local optima. At the same time, benefiting from our instance modification formulation, the neural-based instance modifier can be trained efficiently without any ground-truth supervision, ensuring the framework maintains practicality. Extensive experiments on large-scale TSP benchmarks and real-world benchmarks demonstrate that TSP-MDF significantly improves the performance of traditional heuristics tour constructors, achieving solution quality comparable to neural-based heuristic tour constructors, but with an extremely short training time.", "date": "2026-01-31", "categories": []} {"id": "2602.01090", "title": "Hard Constraints Meet Soft Generation: Guaranteed Feasibility for LLM-based Combinatorial Optimization", "abstract": "Large language models (LLMs) have emerged as promising general-purpose solvers for combinatorial optimization (CO), yet they fundamentally lack mechanisms to guarantee solution feasibility which is critical for real-world deployment. In this work, we introduce FALCON, a framework that ensures 100\\% feasibility through three key innovations: (i) \\emph{grammar-constrained decoding} enforces syntactic validity, (ii) a \\emph{feasibility repair layer} corrects semantic constraint violations, and (iii) \\emph{adaptive Best-of-$N$ sampling} allocates inference compute efficiently. To train the underlying LLM, we introduce the Best-anchored Objective-guided Preference Optimization (BOPO) in LLM training, which weights preference pairs by their objective gap, providing dense supervision without human labels. Theoretically, we prove convergence for BOPO and provide bounds on repair-induced quality loss. Empirically, across seven NP-hard CO problems, FALCON achieves perfect feasibility while matching or exceeding the solution quality of state-of-the-art neural and LLM-based solvers.", "date": "2026-02-01", "categories": []} {"id": "2601.22419", "title": "Dynamic Welfare-Maximizing Pooled Testing", "abstract": "Pooled testing is a common strategy for public health disease screening under limited testing resources, allowing multiple biological samples to be tested together with the resources of a single test, at the cost of reduced individual resolution. While dynamic and adaptive strategies have been extensively studied in the classical pooled testing literature, where the goal is to minimize the number of tests required for full diagnosis of a given population, much of the existing work on welfare-maximizing pooled testing adopts static formulations in which all tests are assigned in advance. In this paper, we study dynamic welfare-maximizing pooled testing strategies in which a limited number of tests are performed sequentially to maximize social welfare, defined as the aggregate utility of individuals who are confirmed to be healthy. We formally define the dynamic problem and study algorithmic approaches for sequential test assignment. Because exact dynamic optimization is computationally infeasible beyond small instances, we evaluate a range of strategies (including exact optimization baselines, greedy heuristics, mixed-integer programming relaxations, and learning-based policies) and empirically characterize their performance and tradeoffs using synthetic experiments. Our results show that dynamic testing can yield substantial welfare improvements over static baselines in low-budget regimes. We find that much of the benefit of dynamic testing is captured by simple greedy policies, which substantially outperform static approaches while remaining computationally efficient. Learning-based methods are included as flexible baselines, but in our experiments they do not reliably improve upon these heuristics. Overall, this work provides a principled computational perspective on dynamic pooled testing and clarifies when dynamic assignment meaningfully improves welfare in public health screening.", "date": "2026-01-30", "categories": []} {"id": "2505.19497", "title": "Learning for Dynamic Combinatorial Optimization without Training Data", "abstract": "We introduce DyCO-GNN, a novel unsupervised learning framework for Dynamic Combinatorial Optimization that requires no training data beyond the problem instance itself. DyCO-GNN leverages structural similarities across time-evolving graph snapshots to accelerate optimization while maintaining solution quality. We evaluate DyCO-GNN on dynamic maximum cut, maximum independent set, and the traveling salesman problem across diverse datasets of varying sizes, demonstrating its superior performance under tight and moderate time budgets. DyCO-GNN consistently outperforms the baseline methods, achieving high-quality solutions up to 3-60x faster, highlighting its practical effectiveness in rapidly evolving resource-constrained settings.", "date": "2025-05-26", "categories": []} {"id": "2506.17357", "title": "Speeding up Local Optimization in Vehicle Routing with Tensor-based GPU Acceleration", "abstract": "Local search plays a central role in many effective heuristic algorithms for the vehicle routing problem (VRP) and its variants. However, neighborhood exploration is known to be computationally expensive and time consuming, especially for large instances or problems with complex constraints. In this study, we explore a promising direction to address this challenge by introducing an original tensor-based GPU acceleration method designed to speed up the commonly used local search operators in vehicle routing. By using an attribute-based representation, the method offers broad extensibility, making it applicable to different VRP variants. Its low-coupling architecture, with intensive computations completely offloaded to the GPU, ensures seamless integration in various local search-based algorithms and frameworks, leading to significant improvements in computational efficiency and potentially improved solution quality. Through comparative experiments on benchmark instances of three routing problems, we demonstrate the substantial computational advantages of the proposed approach over traditional CPU-based implementations. We also provide a detailed analysis of the strengths and limitations of the method, providing valuable insights into its performance characteristics and identifying potential bottlenecks in practical applications. These findings contribute to a better understanding and suggest directions for future improvements.", "date": "2025-06-20", "categories": []} {"id": "2510.01824", "title": "Black-Box Combinatorial Optimization with Order-Invariant Reinforcement Learning", "abstract": "We introduce an order-invariant reinforcement learning framework for black-box combinatorial optimization. Classical estimation-of-distribution algorithms (EDAs) often rely on learning explicit variable dependency graphs, which can be costly and fail to capture complex interactions efficiently. In contrast, we parameterize a multivariate autoregressive generative model trained without a fixed variable ordering. By sampling random generation orders during training, a form of information-preserving dropout, the model is encouraged to be invariant to variable order, promoting search-space diversity, and shaping the model to focus on the most relevant variable dependencies, improving sample efficiency. We adapt Group Relative Policy Optimization (GRPO) to this setting, providing stable policy-gradient updates from scale-invariant advantages. Across a wide range of benchmark algorithms and problem instances of varying sizes, our method frequently achieves the best performance and consistently avoids catastrophic failures.", "date": "2025-10-02", "categories": []} {"id": "2511.10233", "title": "Bridging Synthetic and Real Routing Problems via LLM-Guided Instance Generation and Progressive Adaptation", "abstract": "Recent advances in Neural Combinatorial Optimization (NCO) methods have significantly improved the capability of neural solvers to handle synthetic routing instances. Nonetheless, existing neural solvers typically struggle to generalize effectively from synthetic, uniformly-distributed training data to real-world VRP scenarios, including widely recognized benchmark instances from TSPLib and CVRPLib. To bridge this generalization gap, we present Evolutionary Realistic Instance Synthesis (EvoReal), which leverages an evolutionary module guided by large language models (LLMs) to generate synthetic instances characterized by diverse and realistic structural patterns. Specifically, the evolutionary module produces synthetic instances whose structural attributes statistically mimics those observed in authentic real-world instances. Subsequently, pre-trained NCO models are progressively refined, firstly aligning them with these structurally enriched synthetic distributions and then further adapting them through direct fine-tuning on actual benchmark instances. Extensive experimental evaluations demonstrate that EvoReal markedly improves the generalization capabilities of state-of-the-art neural solvers, yielding a notable reduced performance gap compared to the optimal solutions on the TSPLib (1.05%) and CVRPLib (2.71%) benchmarks across a broad spectrum of problem scales.", "date": "2025-11-13", "categories": []} {"id": "2601.21058", "title": "Snowball: A Scalable All-to-All Ising Machine with Dual-Mode Markov Chain Monte Carlo Spin Selection and Asynchronous Spin Updates for Fast Combinatorial Optimization", "abstract": "Ising machines have emerged as accelerators for combinatorial optimization. To enable practical deployment, this work aims to reduce time-to-solution by addressing three challenges: (1) hardware topology, (2) spin selection and update algorithms, and (3) scalable coupling-coefficient precision. Restricted topologies require minor embedding; naive parallel updates can oscillate or stall; and limited precision can preclude feasible mappings or degrade solution quality.\n This work presents Snowball, a digital, scalable, all-to-all coupled Ising machine that integrates dual-mode Markov chain Monte Carlo spin selection with asynchronous spin updates to promote convergence and reduce time-to-solution. The digital architecture supports wide, configurable coupling precision, unlike many analog realizations at high bit widths. A prototype on an AMD Alveo U250 accelerator card achieves an 8$\\times$ reduction in time-to-solution relative to a state-of-the-art Ising machine on the same benchmark instance.", "date": "2026-01-28", "categories": []} {"id": "2601.21281", "title": "EGAM: Extended Graph Attention Model for Solving Routing Problems", "abstract": "Neural combinatorial optimization (NCO) solvers, implemented with graph neural networks (GNNs), have introduced new approaches for solving routing problems. Trained with reinforcement learning (RL), the state-of-the-art graph attention model (GAM) achieves near-optimal solutions without requiring expert knowledge or labeled data. In this work, we generalize the existing graph attention mechanism and propose the extended graph attention model (EGAM). Our model utilizes multi-head dot-product attention to update both node and edge embeddings, addressing the limitations of the conventional GAM, which considers only node features. We employ an autoregressive encoder-decoder architecture and train it with policy gradient algorithms that incorporate a specially designed baseline. Experiments show that EGAM matches or outperforms existing methods across various routing problems. Notably, the proposed model demonstrates exceptional performance on highly constrained problems, highlighting its efficiency in handling complex graph structures.", "date": "2026-01-29", "categories": []} {"id": "2601.21389", "title": "Learning to Optimize Job Shop Scheduling Under Structural Uncertainty", "abstract": "The Job-Shop Scheduling Problem (JSSP), under various forms of manufacturing uncertainty, has recently attracted considerable research attention. Most existing studies focus on parameter uncertainty, such as variable processing times, and typically adopt the actor-critic framework. In this paper, we explore a different but prevalent form of uncertainty in JSSP: structural uncertainty. Structural uncertainty arises when a job may follow one of several routing paths, and the selection is determined not by policy, but by situational factors (e.g., the quality of intermediate products) that cannot be known in advance. Existing methods struggle to address this challenge due to incorrect credit assignment: a high-quality action may be unfairly penalized if it is followed by a time-consuming path. To address this problem, we propose a novel method named UP-AAC. In contrast to conventional actor-critic methods, UP-AAC employs an asymmetric architecture. While its actor receives a standard stochastic state, the critic is crucially provided with a deterministic state reconstructed in hindsight. This design allows the critic to learn a more accurate value function, which in turn provides a lower-variance policy gradient to the actor, leading to more stable learning. In addition, we design an attention-based Uncertainty Perception Model (UPM) to enhance the actor's scheduling decisions. Extensive experiments demonstrate that our method outperforms existing approaches in reducing makespan on benchmark instances.", "date": "2026-01-29", "categories": []} {"id": "2601.21565", "title": "Multi-objective Integer Linear Programming approach for Automatic Software Cognitive Complexity Reduction", "abstract": "Clear and concise code is necessary to ensure maintainability, so it is crucial that the software is as simple as possible to understand, to avoid bugs and, above all, vulnerabilities. There are many ways to enhance software without changing its functionality, considering the extract method refactoring the primary process to reduce the effort required for code comprehension. The cognitive complexity measure employed in this work is the one defined by SonarSource, which is a company that develops well-known applications for static code analysis. This extraction problem can be modeled as a combinatorial optimization problem. The main difficulty arises from the existence of different criteria for evaluating the solutions obtained, requiring the formulation of the code extraction problem as a multi-objective optimization problem using alternative methods. We propose a multi-objective integer linear programming model to obtain a set of solutions that reduce the cognitive complexity of a given piece of code, such as balancing the number of lines of code and its cognitive complexity. In addition, several algorithms have been developed to validate the model. These algorithms have been integrated into a tool that enables the parameterised resolution of the problem of reducing software cognitive complexity.", "date": "2026-01-29", "categories": []} {"id": "2601.21775", "title": "Differentiable Knapsack and Top-k Operators via Dynamic Programming", "abstract": "Knapsack and Top-k operators are useful for selecting discrete subsets of variables. However, their integration into neural networks is challenging as they are piecewise constant, yielding gradients that are zero almost everywhere. In this paper, we propose a unified framework casting these operators as dynamic programs, and derive differentiable relaxations by smoothing the underlying recursions. On the algorithmic side, we develop efficient parallel algorithms supporting both deterministic and stochastic forward passes, and vector-Jacobian products for the backward pass. On the theoretical side, we prove that Shannon entropy is the unique regularization choice yielding permutation-equivariant operators, and characterize regularizers inducing sparse selections. Finally, on the experimental side, we demonstrate our framework on a decision-focused learning benchmark, a constrained dynamic assortment RL problem, and an extension of discrete VAEs.", "date": "2026-01-29", "categories": []} {"id": "2601.21990", "title": "Batched First-Order Methods for Parallel LP Solving in MIP", "abstract": "We present a batched first-order method for solving multiple linear programs in parallel on GPUs. Our approach extends the primal-dual hybrid gradient algorithm to efficiently solve batches of related linear programming problems that arise in mixed-integer programming techniques such as strong branching and bound tightening. By leveraging matrix-matrix operations instead of repeated matrix-vector operations, we obtain significant computational advantages on GPU architectures. We demonstrate the effectiveness of our approach on various case studies and identify the problem sizes where first-order methods outperform traditional simplex-based solvers depending on the computational environment one can use. This is a significant step for the design and development of integer programming algorithms tightly exploiting GPU capabilities where we argue that some specific operations should be allocated to GPUs and performed in full instead of using light-weight heuristic approaches on CPUs.", "date": "2026-01-29", "categories": []} {"id": "2305.08367", "title": "Toward Highly Efficient and Private Submodular Maximization via Matrix-Based Acceleration", "abstract": "Submodular function maximization is a critical building block for diverse tasks, such as document summarization, sensor placement, and image segmentation. Yet its practical utility is often limit by the $O(knd^2)$ computational bottleneck. In this paper, we propose an integrated framework that addresses efficiency and privacy simultaneously. First, we introduce a novel matrix-based computation paradigm that accelerates function evaluations. Second, we develop approximate data structures that further streamline the optimization process, achieving a theoretical complexity of $O(\\epsilon^{-2}(nd+kn+kd^2)\\log(k/\\delta))$. Third, we integrate ($\\epsilon, \\delta$)-DP guaranties to address the privacy concerns inherent in sensitive optimization tasks.", "date": "2023-05-15", "categories": []} {"id": "2510.04480", "title": "FourierCSP: Differentiable Constraint Satisfaction Problem Solving by Walsh-Fourier Expansion", "abstract": "The Constraint-satisfaction problem (CSP) is fundamental in mathematics, physics, and theoretical computer science. Continuous local search (CLS) solvers, as recent advancements, can achieve highly competitive results on certain classes of Boolean satisfiability (SAT) problems. Motivated by these advances, we extend the CLS framework from Boolean SAT to general CSP with finite-domain variables and expressive constraint formulations. We present FourierCSP, a continuous optimization framework that generalizes the Walsh-Fourier transform to CSP, allowing for transforming versatile constraints to compact multilinear polynomials, thereby avoiding the need for auxiliary variables and memory-intensive encodings. We employ projected subgradient and mirror descent algorithms with provable convergence guarantees, and further combine them to accelerate gradient-based optimization. Empirical results on benchmark suites demonstrate that FourierCSP is scalable and competitive, significantly broadening the class of problems that can be efficiently solved by differentiable CLS techniques and paving the way toward end-to-end neurosymbolic integration.", "date": "2025-10-06", "categories": []} {"id": "2601.20696", "title": "Enterprise Resource Planning Using Multi-type Transformers in Ferro-Titanium Industry", "abstract": "Combinatorial optimization problems such as the Job-Shop Scheduling Problem (JSP) and Knapsack Problem (KP) are fundamental challenges in operations research, logistics, and eterprise resource planning (ERP). These problems often require sophisticated algorithms to achieve near-optimal solutions within practical time constraints. Recent advances in deep learning have introduced transformer-based architectures as promising alternatives to traditional heuristics and metaheuristics. We leverage the Multi-Type Transformer (MTT) architecture to address these benchmarks in a unified framework. We present an extensive experimental evaluation across standard benchmark datasets for JSP and KP, demonstrating that MTT achieves competitive performance on different size of these benchmark problems. We showcase the potential of multi-type attention on a real application in Ferro-Titanium industry. To the best of our knowledge, we are the first to apply multi-type transformers in real manufacturing.", "date": "2026-01-28", "categories": []} {"id": "2601.19041", "title": "HEATACO: Heatmap-Guided Ant Colony Decoding for Large-Scale Travelling Salesman Problems", "abstract": "Heatmap-based non-autoregressive solvers for large-scale Travelling Salesman Problems output dense edge-probability scores, yet final performance largely hinges on the decoder that must satisfy degree-2 constraints and form a single Hamiltonian tour. Greedy commitment can cascade into irreparable mistakes at large $N$, whereas MCTS-guided local search is accurate but compute-heavy and highly engineered. We instead treat the heatmap as a soft edge prior and cast decoding as probabilistic tour construction under feasibility constraints, where the key is to correct local mis-rankings via inexpensive global coordination. Based on this view, we introduce HeatACO, a plug-and-play Max-Min Ant System decoder whose transition policy is softly biased by the heatmap while pheromone updates provide lightweight, instance-specific feedback to resolve global conflicts; optional 2-opt/3-opt post-processing further improves tour quality. On TSP500/1K/10K, using heatmaps produced by four pretrained predictors, HeatACO+2opt achieves gaps down to 0.11%/0.23%/1.15% with seconds-to-minutes CPU decoding for fixed heatmaps, offering a better quality--time trade-off than greedy decoding and published MCTS-based decoders. Finally, we find the gains track heatmap reliability: under distribution shift, miscalibration and confidence collapse bound decoding improvements, suggesting heatmap generalisation is a primary lever for further progress.", "date": "2026-01-26", "categories": []} {"id": "2601.18007", "title": "Memory-Efficient FPGA Implementation of Stochastic Simulated Annealing", "abstract": "Simulated annealing (SA) is a well-known algorithm for solving combinatorial optimization problems. However, the computation time of SA increases rapidly, as the size of the problem grows. Recently, a stochastic simulated annealing (SSA) algorithm that converges faster than conventional SA has been reported. In this paper, we present a hardware-aware SSA (HA- SSA) algorithm for memory-efficient FPGA implementations. HA-SSA can reduce the memory usage of storing intermediate results while maintaining the computing speed of SSA. For evaluation purposes, the proposed algorithm is compared with the conventional SSA and SA approaches on maximum cut combinatorial optimization problems. HA-SSA achieves a convergence speed that is up to 114-times faster than that of the conventional SA algorithm depending on the maximum cut problem selected from the G-set which is a dataset of the maximum cut problems. HA-SSA is implemented on a field-programmable gate array (FPGA) (Xilinx Kintex-7), and it achieves up to 6-times the memory efficiency of conventional SSA while maintaining high solution quality for optimization problems.", "date": "2026-01-25", "categories": []} {"id": "2601.16849", "title": "The Art of Being Difficult: Combining Human and AI Strengths to Find Adversarial Instances for Heuristics", "abstract": "We demonstrate the power of human-LLM collaboration in tackling open problems in theoretical computer science. Focusing on combinatorial optimization, we refine outputs from the FunSearch algorithm [Romera-Paredes et al., Nature 2023] to derive state-of-the-art lower bounds for standard heuristics. Specifically, we target the generation of adversarial instances where these heuristics perform poorly. By iterating on FunSearch's outputs, we identify improved constructions for hierarchical $k$-median clustering, bin packing, the knapsack problem, and a generalization of Lov\\'asz's gasoline problem - some of these have not seen much improvement for over a decade, despite intermittent attention. These results illustrate how expert oversight can effectively extrapolate algorithmic insights from LLM-based evolutionary methods to break long-standing barriers.\n Our findings demonstrate that while LLMs provide critical initial patterns, human expertise is essential for transforming these patterns into mathematically rigorous and insightful constructions. This work highlights that LLMs are a strong collaborative tool in mathematics and computer science research.", "date": "2026-01-23", "categories": []} {"id": "2601.15561", "title": "Enhanced Convergence in p-bit Based Simulated Annealing with Partial Deactivation for Large-Scale Combinatorial Optimization Problems", "abstract": "This article critically investigates the limitations of the simulated annealing algorithm using probabilistic bits (pSA) in solving large-scale combinatorial optimization problems. The study begins with an in-depth analysis of the pSA process, focusing on the issues resulting from unexpected oscillations among p-bits. These oscillations hinder the energy reduction of the Ising model and thus obstruct the successful execution of pSA in complex tasks. Through detailed simulations, we unravel the root cause of this energy stagnation, identifying the feedback mechanism inherent to the pSA operation as the primary contributor to these disruptive oscillations. To address this challenge, we propose two novel algorithms, time average pSA (TApSA) and stalled pSA (SpSA). These algorithms are designed based on partial deactivation of p-bits and are thoroughly tested using Python simulations on maximum cut benchmarks that are typical combinatorial optimization problems. On the 16 benchmarks from 800 to 5,000 nodes, the proposed methods improve the normalized cut value from 0.8% to 98.4% on average in comparison with the conventional pSA.", "date": "2026-01-22", "categories": []} {"id": "2601.15717", "title": "Investigation of the Generalisation Ability of Genetic Programming-evolved Scheduling Rules in Dynamic Flexible Job Shop Scheduling", "abstract": "Dynamic Flexible Job Shop Scheduling (DFJSS) is a complex combinatorial optimisation problem that requires simultaneous machine assignment and operation sequencing decisions in dynamic production environments. Genetic Programming (GP) has been widely applied to automatically evolve scheduling rules for DFJSS. However, existing studies typically train and test GP-evolved rules on DFJSS instances of the same type, which differ only by random seeds rather than by structural characteristics, leaving their cross-type generalisation ability largely unexplored. To address this gap, this paper systematically investigates the generalisation ability of GP-evolved scheduling rules under diverse DFJSS conditions. A series of experiments are conducted across multiple dimensions, including problem scale (i.e., the number of machines and jobs), key job shop parameters (e.g., utilisation level), and data distributions, to analyse how these factors influence GP performance on unseen instance types. The results show that good generalisation occurs when the training instances contain more jobs than the test instances while keeping the number of machines fixed, and when both training and test instances have similar scales or job shop parameters. Further analysis reveals that the number and distribution of decision points in DFJSS instances play a crucial role in explaining these performance differences. Similar decision point distributions lead to better generalisation, whereas significant discrepancies result in a marked degradation of performance. Overall, this study provides new insights into the generalisation ability of GP in DFJSS and highlights the necessity of evolving more generalisable GP rules capable of handling heterogeneous DFJSS instances effectively.", "date": "2026-01-22", "categories": []} {"id": "2601.16056", "title": "Designing faster mixed integer linear programming algorithm via learning the optimal path", "abstract": "Designing faster algorithms for solving Mixed-Integer Linear Programming (MILP) problems is highly desired across numerous practical domains, as a vast array of complex real-world challenges can be effectively modeled as MILP formulations. Solving these problems typically employs the branch-and-bound algorithm, the core of which can be conceived as searching for a path of nodes (or sub-problems) that contains the optimal solution to the original MILP problem. Traditional approaches to finding this path rely heavily on hand-crafted, intuition-based heuristic strategies, which often suffer from unstable and unpredictable performance across different MILP problem instances. To address this limitation, we introduce DeepBound, a deep learning-based node selection algorithm that automates the learning of such human intuition from data. The core of DeepBound lies in learning to prioritize nodes containing the optimal solution, thereby improving solving efficiency. DeepBound introduces a multi-level feature fusion network to capture the node representations. To tackle the inherent node imbalance in branch-and-bound trees, DeepBound employs a pairwise training paradigm that enhances the model's ability to discriminate between nodes. Extensive experiments on three NP-hard MILP benchmarks demonstrate that DeepBound achieves superior solving efficiency over conventional heuristic rules and existing learning-based approaches, obtaining optimal feasible solutions with significantly reduced computation time. Moreover, DeepBound demonstrates strong generalization capability on large and complex instances. The analysis of its learned features reveals that the method can automatically discover more flexible and robust feature selection, which may effectively improve and potentially replace human-designed heuristic rules.", "date": "2026-01-22", "categories": []} {"id": "2601.16194", "title": "A Rolling-Space Branch-and-Price Algorithm for the Multi-Compartment Vehicle Routing Problem with Multiple Time Windows", "abstract": "This paper investigates the multi-compartment vehicle routing problem with multiple time windows (MCVRPMTW), an extension of the classical vehicle routing problem with time windows that considers vehicles equipped with multiple compartments and customers requiring service across several delivery time windows. The problem incorporates three key compartment-related features: (i) compartment flexibility in the number of compartments, (ii) item-to-compartment compatibility, and (iii) item-to-item compatibility. The problem also accommodates practical operational requirements such as driver breaks. To solve the MCVRPMTW, we develop an exact branch-and-price (B&P) algorithm in which the pricing problem is solved using a labeling algorithm. Several acceleration strategies are introduced to limit symmetry during label extensions, improve the stability of dual solutions in column generation, and enhance the branching process. To handle large-scale instances, we propose a rolling-space B&P algorithm that integrates clustering techniques into the solution framework. Extensive computational experiments on instances inspired by a real-world industrial application demonstrate the effectiveness of the proposed approach and provide useful managerial insights for practical implementation.", "date": "2026-01-22", "categories": []} {"id": "2601.14784", "title": "Towards Bound Consistency for the No-Overlap Constraint Using MDDs", "abstract": "Achieving bound consistency for the no-overlap constraint is known to be NP-complete. Therefore, several polynomial-time tightening techniques, such as edge finding, not-first-not-last reasoning, and energetic reasoning, have been introduced for this constraint. In this work, we derive the first bound-consistent algorithm for the no-overlap constraint. By building on the no-overlap MDD defined by Cir\\'e and van Hoeve, we extract bounds of the time window of the jobs, allowing us to tighten start and end times in time polynomial in the number of nodes of the MDD. Similarly, to bound the size and time-complexity, we limit the width of the MDD to a threshold, creating a relaxed MDD that can also be used to relax the bound-consistent filtering. Through experiments on a sequencing problem with time windows and a just-in-time objective ($1 \\mid r_j, d_j, \\bar{d}_j \\mid \\sum E_j + \\sum T_j$), we observe that the proposed filtering, even with a threshold on the width, achieves a stronger reduction in the number of nodes visited in the search tree compared to the previously proposed precedence-detection algorithm of Cir\\'e and van Hoeve. The new filtering also appears to be complementary to classical propagation methods for the no-overlap constraint, allowing a substantial reduction in both the number of nodes and the solving time on several instances.", "date": "2026-01-21", "categories": []} {"id": "2601.15038", "title": "A Curriculum-Based Deep Reinforcement Learning Framework for the Electric Vehicle Routing Problem", "abstract": "The electric vehicle routing problem with time windows (EVRPTW) is a complex optimization problem in sustainable logistics, where routing decisions must minimize total travel distance, fleet size, and battery usage while satisfying strict customer time constraints. Although deep reinforcement learning (DRL) has shown great potential as an alternative to classical heuristics and exact solvers, existing DRL models often struggle to maintain training stability-failing to converge or generalize when constraints are dense. In this study, we propose a curriculum-based deep reinforcement learning (CB-DRL) framework designed to resolve this instability. The framework utilizes a structured three-phase curriculum that gradually increases problem complexity: the agent first learns distance and fleet optimization (Phase A), then battery management (Phase B), and finally the full EVRPTW (Phase C). To ensure stable learning across phases, the framework employs a modified proximal policy optimization algorithm with phase-specific hyperparameters, value and advantage clipping, and adaptive learning-rate scheduling. The policy network is built upon a heterogeneous graph attention encoder enhanced by global-local attention and feature-wise linear modulation. This specialized architecture explicitly captures the distinct properties of depots, customers, and charging stations. Trained exclusively on small instances with N=10 customers, the model demonstrates robust generalization to unseen instances ranging from N=5 to N=100, significantly outperforming standard baselines on medium-scale problems. Experimental results confirm that this curriculum-guided approach achieves high feasibility rates and competitive solution quality on out-of-distribution instances where standard DRL baselines fail, effectively bridging the gap between neural speed and operational reliability.", "date": "2026-01-21", "categories": []} {"id": "2601.15131", "title": "Vehicle Routing with Finite Time Horizon using Deep Reinforcement Learning with Improved Network Embedding", "abstract": "In this paper, we study the vehicle routing problem with a finite time horizon. In this routing problem, the objective is to maximize the number of customer requests served within a finite time horizon. We present a novel routing network embedding module which creates local node embedding vectors and a context-aware global graph representation. The proposed Markov decision process for the vehicle routing problem incorporates the node features, the network adjacency matrix and the edge features as components of the state space. We incorporate the remaining finite time horizon into the network embedding module to provide a proper routing context to the embedding module. We integrate our embedding module with a policy gradient-based deep Reinforcement Learning framework to solve the vehicle routing problem with finite time horizon. We trained and validated our proposed routing method on real-world routing networks, as well as synthetically generated Euclidean networks. Our experimental results show that our method achieves a higher customer service rate than the existing routing methods. Additionally, the solution time of our method is significantly lower than that of the existing methods.", "date": "2026-01-21", "categories": []} {"id": "2401.13770", "title": "AlphaMapleSAT: An MCTS-based Cube-and-Conquer SAT Solver for Hard Combinatorial Problems", "abstract": "This paper introduces AlphaMapleSAT, a Cube-and-Conquer (CnC) parallel SAT solver that integrates Monte Carlo Tree Search (MCTS) with deductive feedback to efficiently solve challenging combinatorial SAT problems. Traditional lookahead cubing methods, used by solvers such as March, limit their search depth to reduce overhead often resulting in suboptimal partitions. By contrast, AlphaMapleSAT performs a deeper MCTS search guided by deductive rewards from SAT solvers. This approach enables informed exploration of the cubing space while keeping cubing costs low. We demonstrate the efficacy of our technique via extensive evaluations against the widely used and established March cubing solver on three well-known challenging combinatorial benchmarks, including the minimum Kochen-Specker (KS) problem from quantum mechanics, the Murty-Simon Conjecture, and the Ramsey problems from extremal graph theory. We compare AlphaMapleSAT against March using different types of conquering solvers such as SAT Modulo Symmetries (SMS) and SAT+CAS, both built on top of the CaDiCaL SAT solver. We show that in all cases, there is a speedup in elapsed real time (wall clock time) ranging from 1.61x to 7.57x on a 128 core machine for the above-mentioned problems. We also perform cube-level and parallel scaling analysis over 32, 64, and 128 cores, which shows that AlphaMapleSAT outperforms March on all these settings. Our results show that deductively-guided MCTS search technique for cubing in CnC solvers can significantly outperform March on hard combinatorial problems.", "date": "2024-01-24", "categories": []} {"id": "2503.06431", "title": "Fairness-aware kidney exchange and kidney paired donation", "abstract": "The kidney paired donation (KPD) program provides an innovative solution to overcome incompatibility challenges in kidney transplants by matching incompatible donor-patient pairs and facilitating kidney exchanges. To address unequal access to transplant opportunities, there are two widely used fairness criteria: group fairness and individual fairness. However, these criteria do not consider protected patient features, which refer to characteristics legally or ethically recognized as needing protection from discrimination, such as race and gender. Motivated by the calibration principle in machine learning, we introduce a new fairness criterion: the matching outcome should be conditionally independent of the protected feature, given the sensitization level. We integrate this fairness criterion as a constraint within the KPD optimization framework and propose a computationally efficient solution using linearization strategies and column-generation methods. Theoretically, we analyze the associated price of fairness using random graph models. Empirically, we compare our fairness criterion with group fairness and individual fairness through both simulations and a real-data example.", "date": "2025-03-09", "categories": []} {"id": "2508.01724", "title": "ReflecSched: Solving Dynamic Flexible Job-Shop Scheduling via LLM-Powered Hierarchical Reflection", "abstract": "The NP-hard Dynamic Flexible Job-Shop Scheduling (DFJSP) problem involves real-time events and complex routing. While traditional rules are efficient but rigid, deep learning is opaque and requires feature engineering. Large Language Models (LLMs) promise adaptive reasoning without this engineering overhead, yet we find their direct application is suboptimal. Baseline LLMs suffer from three key pitfalls: the long-context paradox, where crucial data is underutilized; an underutilization of expert heuristics; and myopic decision-making. To address this, we propose ReflecSched, a framework that empowers the LLM beyond a direct scheduler by equipping it with a strategic analysis capability. ReflecSched tasks the LLM to analyze heuristic-driven simulations across multiple planning horizons and distill them into a concise, natural-language summary termed Strategic Experience. This summary is then integrated into the prompt of a final decision-making module, guiding it to produce non-myopic actions. Experiments demonstrate ReflecSched achieves superior performance, with its best variants attaining an average RPD of 6.09% and rank of 4.39 on GEN-Bench, significantly outperforming strong traditional and learning-based methods including HMPSAC and IDDQN. It also statistically and decisively surpasses direct LLM baselines, securing a 71.35% Win Rate while being, on average, 15.1% more token-efficient on Normal-scale problems. Furthermore, cumulative runtime analysis reveals that ReflecSched's zero-shot nature eliminates the training bottleneck, providing a decisive efficiency advantage in high-variability manufacturing environments. Ablation studies attribute this performance to a robust reflection mechanism that leverages high-quality, contrastive experience. Ultimately, the framework's performance is statistically on par with an oracle-like strategy, showcasing its effectiveness and robustness.", "date": "2025-08-03", "categories": []} {"id": "2512.08601", "title": "Heuristics for Combinatorial Optimization via Value-based Reinforcement Learning: A Unified Framework and Analysis", "abstract": "Since the 1990s, considerable empirical work has been carried out to train statistical models, such as neural networks (NNs), as learned heuristics for combinatorial optimization (CO) problems. When successful, such an approach eliminates the need for experts to design heuristics per problem type. Due to their structure, many hard CO problems are amenable to treatment through reinforcement learning (RL). Indeed, we find a wealth of literature training NNs using value-based, policy gradient, or actor-critic approaches, with promising results, both in terms of empirical optimality gaps and inference runtimes. Nevertheless, there has been a paucity of theoretical work undergirding the use of RL for CO problems. To this end, we introduce a unified framework to model CO problems through Markov decision processes (MDPs) and solve them using RL techniques. We provide easy-to-test assumptions under which CO problems can be formulated as equivalent undiscounted MDPs that provide optimal solutions to the original CO problems. Moreover, we establish conditions under which value-based RL techniques converge to approximate solutions of the CO problem with a guarantee on the associated optimality gap. Our convergence analysis provides: (1) a sufficient rate of increase in batch size and projected gradient descent steps at each RL iteration; (2) the resulting optimality gap in terms of problem parameters and targeted RL accuracy; and (3) the importance of a choice of state-space embedding. Together, our analysis illuminates the success (and limitations) of the celebrated deep Q-learning algorithm in this problem context.", "date": "2025-12-09", "categories": []} {"id": "2509.22023", "title": "Teaching Transformers to Solve Combinatorial Problems through Efficient Trial & Error", "abstract": "Despite their proficiency in various language tasks, Large Language Models (LLMs) struggle with combinatorial problems like Satisfiability, Traveling Salesman Problem, or even basic arithmetic. We address this gap through a novel trial & error approach for solving problems in the class NP, where candidate solutions are iteratively generated and efficiently validated using verifiers. We focus on the paradigmatic task of Sudoku and achieve state-of-the-art accuracy (99%) compared to prior neuro-symbolic approaches. Unlike prior work that used custom architectures, our method employs a vanilla decoder-only Transformer (GPT-2) without external tools or function calling. Our method integrates imitation learning of simple Sudoku rules with an explicit Depth-First Search (DFS) exploration strategy involving informed guessing and backtracking. Moving beyond imitation learning, we seek to minimize the number of guesses until reaching a solution. This is achieved using depth-1 guessing, showing empirically that almost all Sudoku can be solved using the puzzle's rules with at most one guess. We provide a rigorous analysis of this setup formalizing its connection to a contextual variant of Min-Sum Set Cover, a well-studied problem in algorithms and stochastic optimization.", "date": "2025-09-26", "categories": []} {"id": "2601.11189", "title": "Policy-Based Deep Reinforcement Learning Hyperheuristics for Job-Shop Scheduling Problems", "abstract": "This paper proposes a policy-based deep reinforcement learning hyper-heuristic framework for solving the Job Shop Scheduling Problem. The hyper-heuristic agent learns to switch scheduling rules based on the system state dynamically. We extend the hyper-heuristic framework with two key mechanisms. First, action prefiltering restricts decision-making to feasible low-level actions, enabling low-level heuristics to be evaluated independently of environmental constraints and providing an unbiased assessment. Second, a commitment mechanism regulates the frequency of heuristic switching. We investigate the impact of different commitment strategies, from step-wise switching to full-episode commitment, on both training behavior and makespan. Additionally, we compare two action selection strategies at the policy level: deterministic greedy selection and stochastic sampling. Computational experiments on standard JSSP benchmarks demonstrate that the proposed approach outperforms traditional heuristics, metaheuristics, and recent neural network-based scheduling methods", "date": "2026-01-16", "categories": []} {"id": "2601.11414", "title": "New Adaptive Mechanism for Large Neighborhood Search using Dual Actor-Critic", "abstract": "Adaptive Large Neighborhood Search (ALNS) is a widely used heuristic method for solving combinatorial optimization problems. ALNS explores the solution space by iteratively using destroy and repair operators with probabilities, which are adjusted by an adaptive mechanism to find optimal solutions. However, the classic ALNS adaptive mechanism does not consider the interaction between destroy and repair operators when selecting them. To overcome this limitation, this study proposes a novel adaptive mechanism. This mechanism enhances the adaptability of the algorithm through a Dual Actor-Critic (DAC) model, which fully considers the fact that the quality of new solutions is jointly determined by the destroy and repair operators. It effectively utilizes the interaction between these operators during the weight adjustment process, greatly improving the adaptability of the ALNS algorithm. In this mechanism, the destroy and repair processes are modeled as independent Markov Decision Processes to guide the selection of operators more accurately. Furthermore, we use Graph Neural Networks to extract key features from problem instances and perform effective aggregation and normalization to enhance the algorithm's transferability to different sizes and characteristics of problems. Through a series of experiments, we demonstrate that the proposed DAC-ALNS algorithm significantly improves solution efficiency and exhibits excellent transferability.", "date": "2026-01-16", "categories": []} {"id": "2507.20923", "title": "Pareto-Grid-Guided Large Language Models for Fast and High-Quality Heuristics Design in Multi-Objective Combinatorial Optimization", "abstract": "Multi-objective combinatorial optimization problems (MOCOP) frequently arise in practical applications that require the simultaneous optimization of conflicting objectives. Although traditional evolutionary algorithms can be effective, they typically depend on domain knowledge and repeated parameter tuning, limiting flexibility when applied to unseen MOCOP instances. Recently, integration of Large Language Models (LLMs) into evolutionary computation has opened new avenues for automatic heuristic generation, using their advanced language understanding and code synthesis capabilities. Nevertheless, most existing approaches predominantly focus on single-objective tasks, often neglecting key considerations such as runtime efficiency and heuristic diversity in multi-objective settings. To bridge this gap, we introduce Multi-heuristics for MOCOP via Pareto-Grid-guided Evolution of LLMs (MPaGE), a novel enhancement of the Simple Evolutionary Multiobjective Optimization (SEMO) framework that leverages LLMs and Pareto Front Grid (PFG) technique. By partitioning the objective space into grids and retaining top-performing candidates to guide heuristic generation, MPaGE utilizes LLMs to prioritize heuristics with semantically distinct logical structures during variation, thus promoting diversity and mitigating redundancy within the population. Through extensive evaluations, MPaGE demonstrates superior performance over existing LLM-based frameworks, and achieves competitive results to traditional Multi-objective evolutionary algorithms (MOEAs), with significantly faster runtime. Our code is available at: https://github.com/langkhachhoha/MPaGE.", "date": "2025-07-28", "categories": []} {"id": "2510.15583", "title": "Attn-JGNN: Attention Enhanced Join-Graph Neural Networks", "abstract": "We propose an Attention Enhanced Join-Graph Neural Networks(Attn-JGNN) model for solving #SAT problems, which significantly improves the solving accuracy. Inspired by the Iterative Join Graph Propagation (IJGP) algorithm, Attn-JGNN uses tree decomposition to encode the CNF formula into a join-graph, then performs iterative message passing on the join-graph, and finally approximates the model number by learning partition functions. In order to further improve the accuracy of the solution, we apply the attention mechanism in and between clusters of the join-graphs, which makes Attn-JGNN pay more attention to the key variables and clusters in probabilistic inference, and reduces the redundant calculation. Finally, our experiments show that our Attn-JGNN model achieves better results than other neural network methods.", "date": "2025-10-17", "categories": []} {"id": "2510.24039", "title": "Geometric Algorithms for Neural Combinatorial Optimization with Constraints", "abstract": "Self-Supervised Learning (SSL) for Combinatorial Optimization (CO) is an emerging paradigm for solving combinatorial problems using neural networks. In this paper, we address a central challenge of SSL for CO: solving problems with discrete constraints. We design an end-to-end differentiable framework that enables us to solve discrete constrained optimization problems with neural networks. Concretely, we leverage algorithmic techniques from the literature on convex geometry and Carath\\'eodory's theorem to decompose neural network outputs into convex combinations of polytope corners that correspond to feasible sets. This decomposition-based approach enables self-supervised training but also ensures efficient quality-preserving rounding of the neural net output into feasible solutions. Extensive experiments in cardinality-constrained optimization show that our approach can consistently outperform neural baselines. We further provide worked-out examples of how our method can be applied beyond cardinality-constrained problems to a diverse set of combinatorial optimization tasks, including finding independent sets in graphs, and solving matroid-constrained problems.", "date": "2025-10-28", "categories": []} {"id": "2601.09100", "title": "DScheLLM: Enabling Dynamic Scheduling through a Fine-Tuned Dual-System Large language Model", "abstract": "Production scheduling is highly susceptible to dynamic disruptions, such as variations in processing times, machine availability, and unexpected task insertions. Conventional approaches typically rely on event-specific models and explicit analytical formulations, which limits their adaptability and generalization across previously unseen disturbances. To overcome these limitations, this paper proposes DScheLLM, a dynamic scheduling approach that leverages fine-tuned large language models within a dual-system (fast-slow) reasoning architecture to address disturbances of different scales. A unified large language model-based framework is constructed to handle dynamic events, where training datasets for both fast and slow reasoning modes are generated using exact schedules obtained from an operations research solver. The Huawei OpenPangu Embedded-7B model is subsequently fine-tuned under the hybrid reasoning paradigms using LoRA. Experimental evaluations on standard job shop scheduling benchmarks demonstrate that the fast-thinking mode can efficiently generate high-quality schedules and the slow-thinking mode can produce solver-compatible and well-formatted decision inputs. To the best of our knowledge, this work represents one of the earliest studies applying large language models to job shop scheduling in dynamic environments, highlighting their considerable potential for intelligent and adaptive scheduling optimization.", "date": "2026-01-14", "categories": []} {"id": "2601.10583", "title": "Combinatorial Optimization Augmented Machine Learning", "abstract": "Combinatorial optimization augmented machine learning (COAML) has recently emerged as a powerful paradigm for integrating predictive models with combinatorial decision-making. By embedding combinatorial optimization oracles into learning pipelines, COAML enables the construction of policies that are both data-driven and feasibility-preserving, bridging the traditions of machine learning, operations research, and stochastic optimization. This paper provides a comprehensive overview of the state of the art in COAML. We introduce a unifying framework for COAML pipelines, describe their methodological building blocks, and formalize their connection to empirical cost minimization. We then develop a taxonomy of problem settings based on the form of uncertainty and decision structure. Using this taxonomy, we review algorithmic approaches for static and dynamic problems, survey applications across domains such as scheduling, vehicle routing, stochastic programming, and reinforcement learning, and synthesize methodological contributions in terms of empirical cost minimization, imitation learning, and reinforcement learning. Finally, we identify key research frontiers. This survey aims to serve both as a tutorial introduction to the field and as a roadmap for future research at the interface of combinatorial optimization and machine learning.", "date": "2026-01-15", "categories": []} {"id": "2507.20703", "title": "A General Framework for Dynamic MAPF using Multi-Shot ASP and Tunnels", "abstract": "MAPF problem aims to find plans for multiple agents in an environment within a given time, such that the agents do not collide with each other or obstacles. Motivated by the execution and monitoring of these plans, we study Dynamic MAPF (D-MAPF) problem, which allows changes such as agents entering/leaving the environment or obstacles being removed/moved. Considering the requirements of real-world applications in warehouses with the presence of humans, we introduce 1) a general definition for D-MAPF (applicable to variations of D-MAPF), 2) a new framework to solve D-MAPF (utilizing multi-shot computation, and allowing different methods to solve D-MAPF), and 3) a new ASP-based method to solve D-MAPF (combining advantages of replanning and repairing methods, with a novel concept of tunnels to specify where agents can move). We have illustrated the strengths and weaknesses of this method by experimental evaluations, from the perspectives of computational performance and quality of solutions.", "date": "2025-07-28", "categories": []} {"id": "2601.05544", "title": "Buffered AUC maximization for scoring systems via mixed-integer optimization", "abstract": "A scoring system is a linear classifier composed of a small number of explanatory variables, each assigned a small integer coefficient. This system is highly interpretable and allows predictions to be made with simple manual calculations without the need for a calculator. Several previous studies have used mixed-integer optimization (MIO) techniques to develop scoring systems for binary classification; however, they have not focused on directly maximizing AUC (i.e., area under the receiver operating characteristic curve), even though AUC is recognized as an essential evaluation metric for scoring systems. Our goal herein is to establish an effective MIO framework for constructing scoring systems that directly maximize the buffered AUC (bAUC) as the tightest concave lower bound on AUC. Our optimization model is formulated as a mixed-integer linear optimization (MILO) problem that maximizes bAUC subject to a group sparsity constraint for limiting the number of questions in the scoring system. Computational experiments using publicly available real-world datasets demonstrate that our MILO method can build scoring systems with superior AUC values compared to the baseline methods based on regularization and stepwise regression. This research contributes to the advancement of MIO techniques for developing highly interpretable classification models.", "date": "2026-01-09", "categories": []} {"id": "2601.07948", "title": "Reinforcement Learning Methods for Neighborhood Selection in Local Search", "abstract": "Reinforcement learning has recently gained traction as a means to improve combinatorial optimization methods, yet its effectiveness within local search metaheuristics specifically remains comparatively underexamined. In this study, we evaluate a range of reinforcement learning-based neighborhood selection strategies -- multi-armed bandits (upper confidence bound, $\\epsilon$-greedy) and deep reinforcement learning methods (proximal policy optimization, double deep $Q$-network) -- and compare them against multiple baselines across three different problems: the traveling salesman problem, the pickup and delivery problem with time windows, and the car sequencing problem. We show how search-specific characteristics, particularly large variations in cost due to constraint violation penalties, necessitate carefully designed reward functions to provide stable and informative learning signals. Our extensive experiments reveal that algorithm performance varies substantially across problems, although that $\\epsilon$-greedy consistently ranks among the best performers. In contrast, the computational overhead of deep reinforcement learning approaches only makes them competitive with a substantially longer runtime. These findings highlight both the promise and the practical limitations of deep reinforcement learning in local search.", "date": "2026-01-12", "categories": []} {"id": "2601.08696", "title": "Enabling Population-Based Architectures for Neural Combinatorial Optimization", "abstract": "Neural Combinatorial Optimization (NCO) has mostly focused on learning policies, typically neural networks, that operate on a single candidate solution at a time, either by constructing one from scratch or iteratively improving it. In contrast, decades of work in metaheuristics have shown that maintaining and evolving populations of solutions improves robustness and exploration, and often leads to stronger performance. To close this gap, we study how to make NCO explicitly population-based by learning policies that act on sets of candidate solutions. We first propose a simple taxonomy of population awareness levels and use it to highlight two key design challenges: (i) how to represent a whole population inside a neural network, and (ii) how to learn population dynamics that balance intensification (generating good solutions) and diversification (maintaining variety). We make these ideas concrete with two complementary tools: one that improves existing solutions using information shared across the whole population, and the other generates new candidate solutions that explicitly balance being high-quality with diversity. Experimental results on Maximum Cut and Maximum Independent Set indicate that incorporating population structure is advantageous for learned optimization methods and opens new connections between NCO and classical population-based search.", "date": "2026-01-13", "categories": []} {"id": "2601.08753", "title": "Grid-Aware Charging and Operational Optimization for Mixed-Fleet Public Transit", "abstract": "The rapid growth of urban populations and the increasing need for sustainable transportation solutions have prompted a shift towards electric buses in public transit systems. However, the effective management of mixed fleets consisting of both electric and diesel buses poses significant operational challenges. One major challenge is coping with dynamic electricity pricing, where charging costs vary throughout the day. Transit agencies must optimize charging assignments in response to such dynamism while accounting for secondary considerations such as seating constraints. This paper presents a comprehensive mixed-integer linear programming (MILP) model to address these challenges by jointly optimizing charging schedules and trip assignments for mixed (electric and diesel bus) fleets while considering factors such as dynamic electricity pricing, vehicle capacity, and route constraints. We address the potential computational intractability of the MILP formulation, which can arise even with relatively small fleets, by employing a hierarchical approach tailored to the fleet composition. By using real-world data from the city of Chattanooga, Tennessee, USA, we show that our approach can result in significant savings in the operating costs of the mixed transit fleets.", "date": "2026-01-13", "categories": []} {"id": "2411.04293", "title": "A Random-Key Optimizer for Combinatorial Optimization", "abstract": "This paper introduces the Random-Key Optimizer (RKO), a versatile and efficient stochastic local search method tailored for combinatorial optimization problems. Using the random-key concept, RKO encodes solutions as vectors of random keys that are subsequently decoded into feasible solutions via problem-specific decoders. The RKO framework is able to combine a plethora of classic metaheuristics, each capable of operating independently or in parallel, with solution sharing facilitated through an elite solution pool. This modular approach allows for the adaptation of various metaheuristics, including simulated annealing, iterated local search, and greedy randomized adaptive search procedures, among others. The efficacy of the RKO framework, implemented in C++ and publicly available (Github public repository: github.com/RKO-solver), is demonstrated through its application to three NP-hard combinatorial optimization problems: the alpha-neighborhood p-median problem, the tree of hubs location problem, and the node-capacitated graph partitioning problem. The results highlight the framework's ability to produce high-quality solutions across diverse problem domains, underscoring its potential as a robust tool for combinatorial optimization.", "date": "2024-11-06", "categories": []} {"id": "2508.03030", "title": "Collab-Solver: Collaborative Solving Policy Learning for Mixed-Integer Linear Programming", "abstract": "Mixed-integer linear programming (MILP) has been a fundamental problem in combinatorial optimization. Conventional MILP solving mainly relies on carefully designed heuristics embedded in the branch-and-bound framework. Driven by the strong capabilities of neural networks, recent research is exploring the value of machine learning alongside conventional MILP solving. Although learning-based MILP methods have shown great promise, existing works typically learn policies for individual modules in MILP solvers in isolation, without considering their interdependence, which limits both solving efficiency and solution quality. To address this limitation, we propose Collab-Solver, a novel multi-agent-based policy learning framework for MILP that enables collaborative policy optimization for multiple modules. Specifically, we formulate the collaboration between cut selection and branching in MILP solving as a Stackelberg game. Under this formulation, we develop a two-phase learning paradigm to stabilize collaborative policy learning: the first phase performs data-communicated policy pretraining, and the second phase further orchestrates the policy learning for various modules. Extensive experiments on both synthetic and large-scale real-world MILP datasets demonstrate that the jointly learned policies significantly improve solving performance. Moreover, the policies learned by Collab-Solver have also demonstrated excellent generalization abilities across different instance sets.", "date": "2025-08-05", "categories": []} {"id": "2508.06899", "title": "GDBA Revisited: Unleashing the Power of Guided Local Search for Distributed Constraint Optimization", "abstract": "Local search is an important class of incomplete algorithms for solving Distributed Constraint Optimization Problems (DCOPs) but it often converges to poor local optima. While Generalized Distributed Breakout Algorithm (GDBA) provides a comprehensive rule set to escape premature convergence, its empirical benefits remain marginal on general-valued problems. In this work, we systematically examine GDBA and identify three factors that potentially lead to its inferior performance, i.e., over-aggressive constraint violation conditions, unbounded penalty accumulation, and uncoordinated penalty updates. To address these issues, we propose Distributed Guided Local Search (DGLS), a novel GLS framework for DCOPs that incorporates an adaptive violation condition to selectively penalize constraints with high cost, a penalty evaporation mechanism to control the magnitude of penalization, and a synchronization scheme for coordinated penalty updates. We theoretically show that the penalty values are bounded, and agents play a potential game in DGLS. Extensive empirical results on various benchmarks demonstrate the great superiority of DGLS over state-of-the-art baselines. Compared to Damped Max-sum with high damping factors, our DGLS achieves competitive performance on general-valued problems, and outperforms by significant margins on structured problems in terms of anytime results.", "date": "2025-08-09", "categories": []} {"id": "2601.06029", "title": "A Recommendation System-Based Framework for Enhancing Human-Machine Collaboration in Industrial Timetabling Rescheduling: Application in Preventive Maintenance", "abstract": "Industrial timetabling is a critical task for decision-makers across various sectors to ensure efficient system operation. In real-world settings, it remains challenging because unexpected events often disrupt execution. When such events arise, effective rescheduling and collaboration between humans and machines becomes essential. This paper presents a recommendation system-based framework for handling rescheduling challenges, built on Timefold, a powerful AI-driven planning engine. Our experimental study evaluates nine instances inspired by a realworld preventive maintenance use case, aiming to identify the heuristic that best balances solution quality and computing time to support near-optimal decisionmaking when rescheduling is required due to unexpected events during operational days. Finally, we illustrate the complete process of our recommendation system through a simple use case.", "date": "2025-11-04", "categories": []} {"id": "2601.06351", "title": "A Fast and Effective Method for Euclidean Anticlustering: The Assignment-Based-Anticlustering Algorithm", "abstract": "The anticlustering problem is to partition a set of objects into K equal-sized anticlusters such that the sum of distances within anticlusters is maximized. The anticlustering problem is NP-hard. We focus on anticlustering in Euclidean spaces, where the input data is tabular and each object is represented as a D-dimensional feature vector. Distances are measured as squared Euclidean distances between the respective vectors. Applications of Euclidean anticlustering include social studies, particularly in psychology, K-fold cross-validation in which each fold should be a good representative of the entire dataset, the creation of mini-batches for gradient descent in neural network training, and balanced K-cut partitioning. In particular, machine-learning applications involve million-scale datasets and very large values of K, making scalable anticlustering algorithms essential. Existing algorithms are either exact methods that can solve only small instances or heuristic methods, among which the most scalable is the exchange-based heuristic fast_anticlustering. We propose a new algorithm, the Assignment-Based Anticlustering algorithm (ABA), which scales to very large instances. A computational study shows that ABA outperforms fast_anticlustering in both solution quality and running time. Moreover, ABA scales to instances with millions of objects and hundreds of thousands of anticlusters within short running times, beyond what fast_anticlustering can handle. As a balanced K-cut partitioning method for tabular data, ABA is superior to the well-known METIS method in both solution quality and running time. The code of the ABA algorithm is available on GitHub.", "date": "2026-01-09", "categories": []} {"id": "2601.06542", "title": "Resource-constrained Project Scheduling with Time-of-Use Energy Tariffs and Machine States: A Logic-based Benders Decomposition Approach", "abstract": "In this paper, we investigate the Resource-Constrained Project Scheduling Problem (RCPSP) with time-of-use energy tariffs (TOU) and machine states, a variant of RCPSP for production scheduling where energy price is part of the criteria and one machine is highly energy-demanding and can be in one of the following three states: proc, idle, or off. The problem involves scheduling all tasks, respecting precedence constraints and resource limitations, while minimizing the combination of the overall makespan and the total energy cost (TEC), which varies according to the TOU pricing, which can take negative values. We propose two novel approaches to solve it: a monolithic Constraint Programming (CP) approach and a Logic-Based Benders Decomposition (LBBD) approach. The latter combines a master problem dealing with energy cost solved using Integer Linear Programming (ILP) with a subproblem handling the RCPSP resolved using CP. Both approaches surpass the monolithic compact ILP approach, but the LBBD significantly outperforms the CP when the ratio of energy-intensive tasks over the overall tasks is moderate, allowing for solving instances with up to 1600 tasks in sparse instances. Finally, we put forth a way of generalizing our LBBD approach to other problems sharing similar characteristics, and we applied it to a problem based on an RCPSP problem with blocking times & total weighted tardiness criterion and a flexible job shop.", "date": "2026-01-10", "categories": []} {"id": "2601.06670", "title": "Otimizando A Aloca\\c{c}\\~ao De Salas De Aula Com Foco Na Acessibilidade Para Pessoas Com Defici\\^encia", "abstract": "This paper addresses the challenge of classroom allocation in higher education institutions, with an explicit emphasis on accessibility for Persons with Disabilities (PwDs). Employing a case study of a university's computer science department, the paper proposes an Integer Linear Programming (ILP)-based optimization model, which is solved using the Gurobi solver. The objective is to minimize the number of classrooms used by prioritizing the assignment of PwD students to ground-floor classrooms to reduce accessibility barriers. The model is calibrated with a weighting parameter, alpha, that allows for a balance between spatial efficiency and promoting accessibility. Experimental results indicate that adjusting alpha can achieve a balance point that significantly improves current manual allocation practices, reducing the number of classrooms required and accessibility penalties. The findings suggest that optimization methods can improve operational efficiency in academic institutions while promoting a more inclusive environment for all students. Future work may expand the application of the model to other departments and contexts and integrate additional criteria to develop a more holistic approach.", "date": "2026-01-10", "categories": []} {"id": "2601.07514", "title": "Data-Driven Stochastic VRP: Integration of Forecast Duration into Optimization for Utility Workforce Management", "abstract": "This paper investigates the integration of machine learning forecasts of intervention durations into a stochastic variant of the Capacitated Vehicle Routing Problem with Time Windows (CVRPTW). In particular, we exploit tree-based gradient boosting (XGBoost) trained on eight years of gas meter maintenance data to produce point predictions and uncertainty estimates, which then drive a multi-objective evolutionary optimization routine. The methodology addresses uncertainty through sub-Gaussian concentration bounds for route-level risk buffers and explicitly accounts for competing operational KPIs through a multi-objective formulation. Empirical analysis of prediction residuals validates the sub-Gaussian assumption underlying the risk model. From an empirical point of view, our results report improvements around 20-25\\% in operator utilization and completion rates compared with plans computed using default durations. The integration of uncertainty quantification and risk-aware optimization provides a practical framework for handling stochastic service durations in real-world routing applications.", "date": "2026-01-12", "categories": []} {"id": "2601.07597", "title": "Pheromone-Focused Ant Colony Optimization algorithm for path planning", "abstract": "Ant Colony Optimization (ACO) is a prominent swarm intelligence algorithm extensively applied to path planning. However, traditional ACO methods often exhibit shortcomings, such as blind search behavior and slow convergence within complex environments. To address these challenges, this paper proposes the Pheromone-Focused Ant Colony Optimization (PFACO) algorithm, which introduces three key strategies to enhance the problem-solving ability of the ant colony. First, the initial pheromone distribution is concentrated in more promising regions based on the Euclidean distances of nodes to the start and end points, balancing the trade-off between exploration and exploitation. Second, promising solutions are reinforced during colony iterations to intensify pheromone deposition along high-quality paths, accelerating convergence while maintaining solution diversity. Third, a forward-looking mechanism is implemented to penalize redundant path turns, promoting smoother and more efficient solutions. These strategies collectively produce the focused pheromones to guide the ant colony's search, which enhances the global optimization capabilities of the PFACO algorithm, significantly improving convergence speed and solution quality across diverse optimization problems. The experimental results demonstrate that PFACO consistently outperforms comparative ACO algorithms in terms of convergence speed and solution quality.", "date": "2026-01-12", "categories": []} {"id": "2601.01132", "title": "Generating Diverse TSP Tours via a Combination of Graph Pointer Network and Dispersion", "abstract": "We address the Diverse Traveling Salesman Problem (D-TSP), a bi-criteria optimization challenge that seeks a set of $k$ distinct TSP tours. The objective requires every selected tour to have a length at most $c|T^*|$ (where $|T^*|$ is the optimal tour length) while minimizing the average Jaccard similarity across all tour pairs. This formulation is crucial for applications requiring both high solution quality and fault tolerance, such as logistics planning, robotics pathfinding or strategic patrolling. Current methods are limited: traditional heuristics, such as the Niching Memetic Algorithm (NMA) or bi-criteria optimization, incur high computational complexity $O(n^3)$, while modern neural approaches (e.g., RF-MA3S) achieve limited diversity quality and rely on complex, external mechanisms.\n To overcome these limitations, we propose a novel hybrid framework that decomposes D-TSP into two efficient steps. First, we utilize a simple Graph Pointer Network (GPN), augmented with an approximated sequence entropy loss, to efficiently sample a large, diverse pool of high-quality tours. This simple modification effectively controls the quality-diversity trade-off without complex external mechanisms. Second, we apply a greedy algorithm that yields a 2-approximation for the dispersion problem to select the final $k$ maximally diverse tours from the generated pool. Our results demonstrate state-of-the-art performance. On the Berlin instance, our model achieves an average Jaccard index of $0.015$, significantly outperforming NMA ($0.081$) and RF-MA3S. By leveraging GPU acceleration, our GPN structure achieves a near-linear empirical runtime growth of $O(n)$. While maintaining solution diversity comparable to complex bi-criteria algorithms, our approach is over 360 times faster on large-scale instances (783 cities), delivering high-quality TSP solutions with unprecedented efficiency and simplicity.", "date": "2026-01-03", "categories": []} {"id": "2601.03646", "title": "ReLA: Representation Learning and Aggregation for Job Scheduling with Reinforcement Learning", "abstract": "Job scheduling is widely used in real-world manufacturing systems to assign ordered job operations to machines under various constraints. Existing solutions remain limited by long running time or insufficient schedule quality, especially when problem scale increases. In this paper, we propose ReLA, a reinforcement-learning (RL) scheduler built on structured representation learning and aggregation. ReLA first learns diverse representations from scheduling entities, including job operations and machines, using two intra-entity learning modules with self-attention and convolution and one inter-entity learning module with cross-attention. These modules are applied in a multi-scale architecture, and their outputs are aggregated to support RL decision-making. Across experiments on small, medium, and large job instances, ReLA achieves the best makespan in most tested settings over the latest solutions. On non-large instances, ReLA reduces the optimality gap of the SOTA baseline by 13.0%, while on large-scale instances it reduces the gap by 78.6%, with the average optimality gaps lowered to 7.3% and 2.1%, respectively. These results confirm that ReLA's learned representations and aggregation provide strong decision support for RL scheduling, and enable fast job completion and decision-making for real-world applications.", "date": "2026-01-07", "categories": []} {"id": "2601.04239", "title": "Solving Cyclic Antibandwidth Problem by SAT", "abstract": "The Cyclic Antibandwidth Problem (CABP), a variant of the Antibandwidth Problem, is an NP-hard graph labeling problem with numerous applications. Despite significant research efforts, existing state-of-the-art approaches for CABP are exclusively heuristic or metaheuristic in nature, and exact methods have been limited to restricted graph classes. In this paper, we present the first exact approach for the CABP on general graphs, based on SAT solving, called SAT-CAB. The proposed method is able to systematically explore the solution space and guarantee global optimality, overcoming the limitations of previously reported heuristic algorithms. This approach relies on a novel and efficient SAT encoding of CABP, in which the problem is transformed into a sequence of At-Most-One constraints. In particular, we introduce a compact representation of the At-Most-One constraints inherent to CABP, which significantly reduces the size of the resulting formulas and enables modern SAT solvers to effectively explore the solution space and to certify global optimality. Extensive computational experiments on standard benchmark instances show that the proposed method efficiently solves CABP instances of practical relevance, while identifying several previously unknown optimal solutions. Moreover, global optimal cyclic antibandwidth values are proven for a number of benchmark instances for the first time. Comparative results indicate that SAT-CAB consistently matches or surpasses the best-known solutions obtained by state-of-the-art heuristic algorithms such as MS-GVNS, HABC-CAB, and MACAB, as well as strong commercial Constraint Programming and Mixed Integer Programming solvers like CPLEX and Gurobi, particularly on general graphs, while also providing optimality guarantees. These results advance the state of the art for CABP and provide a new baseline for exact and hybrid methods on general graphs.", "date": "2026-01-05", "categories": []} {"id": "2601.04509", "title": "A General Neural Backbone for Mixed-Integer Linear Optimization via Dual Attention", "abstract": "Mixed-integer linear programming (MILP), a widely used modeling framework for combinatorial optimization, are central to many scientific and engineering applications, yet remains computationally challenging at scale. Recent advances in deep learning address this challenge by representing MILP instances as variable-constraint bipartite graphs and applying graph neural networks (GNNs) to extract latent structural patterns and enhance solver efficiency. However, this architecture is inherently limited by the local-oriented mechanism, leading to restricted representation power and hindering neural approaches for MILP. Here we present an attention-driven neural architecture that learns expressive representations beyond the pure graph view. A dual-attention mechanism is designed to perform parallel self- and cross-attention over variables and constraints, enabling global information exchange and deeper representation learning. We apply this general backbone to various downstream tasks at the instance level, element level, and solving state level. Extensive experiments across widely used benchmarks show consistent improvements of our approach over state-of-the-art baselines, highlighting attention-based neural architectures as a powerful foundation for learning-enhanced mixed-integer linear optimization.", "date": "2026-01-08", "categories": []} {"id": "2601.04705", "title": "A zone-based training approach for last-mile routing using Graph Neural Networks and Pointer Networks", "abstract": "Rapid e-commerce growth has pushed last-mile delivery networks to their limits, where small routing gains translate into lower costs, faster service, and fewer emissions. Classical heuristics struggle to adapt when travel times are highly asymmetric (e.g., one-way streets, congestion). A deep learning-based approach to the last-mile routing problem is presented to generate geographical zones composed of stop sequences to minimize last-mile delivery times.\n The presented approach is an encoder-decoder architecture. Each route is represented as a complete directed graph whose nodes are stops and whose edge weights are asymmetric travel times. A Graph Neural Network encoder produces node embeddings that captures the spatial relationships between stops. A Pointer Network decoder then takes the embeddings and the route's start node to sequentially select the next stops, assigning a probability to each unvisited node as the next destination.\n Cells of a Discrete Global Grid System which contain route stops in the training data are obtained and clustered to generate geographical zones of similar size in which the process of training and inference are divided. Subsequently, a different instance of the model is trained per zone only considering the stops of the training routes which are included in that zone.\n This approach is evaluated using the Los Angeles routes from the 2021 Amazon Last Mile Routing Challenge. Results from general and zone-based training are compared, showing a reduction in the average predicted route length in the zone-based training compared to the general training. The performance improvement of the zone-based approach becomes more pronounced as the number of stops per route increases.", "date": "2026-01-08", "categories": []} {"id": "2601.04887", "title": "Flexible Manufacturing Systems Intralogistics: Dynamic Optimization of AGVs and Tool Sharing Using Coloured-Timed Petri Nets and Actor-Critic RL with Actions Masking", "abstract": "Flexible Manufacturing Systems (FMS) are pivotal in optimizing production processes in today's rapidly evolving manufacturing landscape. This paper advances the traditional job shop scheduling problem by incorporating additional complexities through the simultaneous integration of automated guided vehicles (AGVs) and tool-sharing systems. We propose a novel approach that combines Colored-Timed Petri Nets (CTPNs) with actor-critic model-based reinforcement learning (MBRL), effectively addressing the multifaceted challenges associated with FMS. CTPNs provide a formal modeling structure and dynamic action masking, significantly reducing the action search space, while MBRL ensures adaptability to changing environments through the learned policy. Leveraging the advantages of MBRL, we incorporate a lookahead strategy for optimal positioning of AGVs, improving operational efficiency. Our approach was evaluated on small-sized public benchmarks and a newly developed large-scale benchmark inspired by the Taillard benchmark. The results show that our approach matches traditional methods on smaller instances and outperforms them on larger ones in terms of makespan while achieving a tenfold reduction in computation time. To ensure reproducibility, we propose a gym-compatible environment and an instance generator. Additionally, an ablation study evaluates the contribution of each framework component to its overall performance.", "date": "2026-01-08", "categories": []} {"id": "2506.13600", "title": "The ASP-based Nurse Scheduling System at the University of Yamanashi Hospital", "abstract": "We present the design principles of a nurse scheduling system built using Answer Set Programming (ASP) and successfully deployed at the University of Yamanashi Hospital. Nurse scheduling is a complex optimization problem requiring the reconciliation of individual nurse preferences with hospital staffing needs across various wards. This involves balancing hard and soft constraints and the flexibility of interactive adjustments. While extensively studied in academia, real-world nurse scheduling presents unique challenges that go beyond typical benchmark problems and competitions. This paper details the practical application of ASP to address these challenges at the University of Yamanashi Hospital, focusing on the insights gained and the advancements in ASP technology necessary to effectively manage the complexities of real-world deployment.\n", "date": "2025-06-16", "categories": []} {"id": "2508.09395", "title": "Tightening the mixed integer linear formulation for the piecewise linear approximation in general dimensions", "abstract": "This paper addresses the problem of tightening the mixed-integer linear programming (MILP) formulation for continuous piecewise linear (CPWL) approximations of data sets in arbitrary dimensions. The MILP formulation leverages the difference-of-convex (DC) representation of CPWL functions. We introduce the concept of well-behaved CPWL interpolations and demonstrate that any CPWL interpolation of a data set has a well-behaved version. This result is critical to tighten the MILP problem. We present six different strategies to tighten the problem, which include fixing the values of some variables, introducing additional constraints, identifying small big-M parameter values and applying tighter variable bounds. These methods leverage key aspects of the DC representation and the inherent structure of well-behaved CPWL interpolations. Experimental results demonstrate that specific combinations of these tightening strategies lead to significant improvement in solution times, especially for tightening strategies that consider well-behaved CPWL solutions.", "date": "2025-08-13", "categories": []} {"id": "2504.02644", "title": "Solving the Paint Shop Problem with Flexible Management of Multi-Lane Buffers Using Reinforcement Learning and Action Masking", "abstract": "In the paint shop problem, an unordered incoming sequence of cars assigned to different colors has to be reshuffled with the objective of minimizing the number of color changes. To reshuffle the incoming sequence, manufacturers can employ a first-in-first-out multi-lane buffer system allowing store and retrieve operations. So far, prior studies primarily focused on simple decision heuristics like greedy or simplified problem variants that do not allow full flexibility when performing store and retrieve operations. In this study, we propose a reinforcement learning approach to minimize color changes for the flexible problem variant, where store and retrieve operations can be performed in an arbitrary order. After proving that greedy retrieval is optimal, we incorporate this finding into the model using action masking. Our evaluation, based on 170 problem instances with 2-8 buffer lanes and 5-15 colors, shows that our approach reduces color changes compared to existing methods by considerable margins depending on the problem size. Furthermore, we demonstrate the robustness of our approach towards different buffer sizes and imbalanced color distributions.", "date": "2025-04-03", "categories": []} {"id": "2601.00023", "title": "A multi-algorithm approach for operational human resources workload balancing in a last mile urban delivery system", "abstract": "Efficient workload assignment to the workforce is critical in last-mile package delivery systems. In this context, traditional methods of assigning package deliveries to workers based on geographical proximity can be inefficient and surely guide to an unbalanced workload distribution among delivery workers. In this paper, we look at the problem of operational human resources workload balancing in last-mile urban package delivery systems. The idea is to consider the effort workload to optimize the system, i.e., the optimization process is now focused on improving the delivery time, so that the workload balancing is complete among all the staff. This process should correct significant decompensations in workload among delivery workers in a given zone. Specifically, we propose a multi-algorithm approach to tackle this problem. The proposed approach takes as input a set of delivery points and a defined number of workers, and then assigns packages to workers, in such a way that it ensures that each worker completes a similar amount of work per day. The proposed algorithms use a combination of distance and workload considerations to optimize the allocation of packages to workers. In this sense, the distance between the delivery points and the location of each worker is also taken into account. The proposed multi-algorithm methodology includes different versions of k-means, evolutionary approaches, recursive assignments based on k-means initialization with different problem encodings, and a hybrid evolutionary ensemble algorithm. We have illustrated the performance of the proposed approach in a real-world problem in an urban last-mile package delivery workforce operating at Azuqueca de Henares, Spain.", "date": "2025-12-24", "categories": []} {"id": "2601.00580", "title": "Priority-Aware Multi-Robot Coverage Path Planning", "abstract": "Multi-robot systems are widely used for coverage tasks that require efficient coordination across large environments. In Multi-Robot Coverage Path Planning (MCPP), the objective is typically to minimize the makespan by generating non-overlapping paths for full-area coverage. However, most existing methods assume uniform importance across regions, limiting their effectiveness in scenarios where some zones require faster attention. We introduce the Priority-Aware MCPP (PA-MCPP) problem, where a subset of the environment is designated as prioritized zones with associated weights. The goal is to minimize, in lexicographic order, the total priority-weighted latency of zone coverage and the overall makespan. To address this, we propose a scalable two-phase framework combining (1) greedy zone assignment with local search, spanning-tree-based path planning, and (2) Steiner-tree-guided residual coverage. Experiments across diverse scenarios demonstrate that our method significantly reduces priority-weighted latency compared to standard MCPP baselines, while maintaining competitive makespan. Sensitivity analyses further show that the method scales well with the number of robots and that zone coverage behavior can be effectively controlled by adjusting priority weights.", "date": "2026-01-02", "categories": []} {"id": "2601.00611", "title": "Stronger Approximation Guarantees for Non-Monotone {\\gamma}-Weakly DR-Submodular Maximization", "abstract": "Maximizing submodular objectives under constraints is a fundamental problem in machine learning and optimization. We study the maximization of a nonnegative, non-monotone $\\gamma$-weakly DR-submodular function over a down-closed convex body. Our main result is an approximation algorithm whose guarantee depends smoothly on $\\gamma$; in particular, when $\\gamma=1$ (the DR-submodular case) our bound recovers the $0.401$ approximation factor, while for $\\gamma<1$ the guarantee degrades gracefully and, it improves upon previously reported bounds for $\\gamma$-weakly DR-submodular maximization under the same constraints. Our approach combines a Frank-Wolfe-guided continuous-greedy framework with a $\\gamma$-aware double-greedy step, yielding a simple yet effective procedure for handling non-monotonicity. This results in state-of-the-art guarantees for non-monotone $\\gamma$-weakly DR-submodular maximization over down-closed convex bodies.", "date": "2026-01-02", "categories": []} {"id": "2601.00689", "title": "Cost Optimization in Production Line Using Genetic Algorithm", "abstract": "This paper presents a genetic algorithm (GA) approach to cost-optimal task scheduling in a production line. The system consists of a set of serial processing tasks, each with a given duration, unit execution cost, and precedence constraints, which must be assigned to an unlimited number of stations subject to a per-station duration bound. The objective is to minimize the total production cost, modeled as a station-wise function of task costs and the duration bound, while strictly satisfying all prerequisite and capacity constraints. Two chromosome encoding strategies are investigated: a station-based representation implemented using the JGAP library with SuperGene validity checks, and a task-based representation in which genes encode station assignments directly. For each encoding, standard GA operators (crossover, mutation, selection, and replacement) are adapted to preserve feasibility and drive the population toward lower-cost schedules. Experimental results on three classes of precedence structures-tightly coupled, loosely coupled, and uncoupled-demonstrate that the task-based encoding yields smoother convergence and more reliable cost minimization than the station-based encoding, particularly when the number of valid schedules is large. The study highlights the advantages of GA over gradient-based and analytical methods for combinatorial scheduling problems, especially in the presence of complex constraints and non-differentiable cost landscapes.", "date": "2026-01-02", "categories": []} {"id": "2601.00770", "title": "LLM Agents for Combinatorial Efficient Frontiers: Investment Portfolio Optimization", "abstract": "Investment portfolio optimization is a task conducted in all major financial institutions. The Cardinality Constrained Mean-Variance Portfolio Optimization (CCPO) problem formulation is ubiquitous for portfolio optimization. The challenge of this type of portfolio optimization, a mixed-integer quadratic programming (MIQP) problem, arises from the intractability of solutions from exact solvers, where heuristic algorithms are used to find approximate portfolio solutions. CCPO entails many laborious and complex workflows and also requires extensive effort pertaining to heuristic algorithm development, where the combination of pooled heuristic solutions results in improved efficient frontiers. Hence, common approaches are to develop many heuristic algorithms. Agentic frameworks emerge as a promising candidate for many problems within combinatorial optimization, as they have been shown to be equally efficient with regard to automating large workflows and have been shown to be excellent in terms of algorithm development, sometimes surpassing human-level performance. This study implements a novel agentic framework for the CCPO and explores several concrete architectures. In benchmark problems, the implemented agentic framework matches state-of-the-art algorithms. Furthermore, complex workflows and algorithm development efforts are alleviated, while in the worst case, lower but acceptable error is reported.", "date": "2026-01-02", "categories": []} {"id": "2510.21128", "title": "A Unified Approach to Submodular Maximization Under Noise", "abstract": "We consider the problem of maximizing a submodular function with access to a noisy value oracle for the function instead of an exact value oracle. Similar to prior work, we assume that the noisy oracle is persistent in that multiple calls to the oracle for a specific set always return the same value. In this model, Hassidim and Singer (2017) design a $(1-1/e)$-approximation algorithm for monotone submodular maximization subject to a cardinality constraint, and Huang et al (2022) design a $(1-1/e)/2$-approximation algorithm for monotone submodular maximization subject to any arbitrary matroid constraint. In this paper, we design a meta-algorithm that allows us to take any \"robust\" algorithm for exact submodular maximization as a black box and transform it into an algorithm for the noisy setting while retaining the approximation guarantee. By using the meta-algorithm with the measured continuous greedy algorithm, we obtain a $(1-1/e)$-approximation (resp. $1/e$-approximation) for monotone (resp. non-monotone) submodular maximization subject to a matroid constraint under noise. Furthermore, by using the meta-algorithm with the double greedy algorithm, we obtain a $1/2$-approximation for unconstrained (non-monotone) submodular maximization under noise.", "date": "2025-10-24", "categories": []} {"id": "2512.24037", "title": "Kidney Exchange: Faster Parameterized Algorithms and Tighter Lower Bounds", "abstract": "The kidney exchange mechanism allows many patient-donor pairs who are otherwise incompatible with each other to come together and exchange kidneys along a cycle. However, due to infrastructure and legal constraints, kidney exchange can only be performed in small cycles in practice. In reality, there are also some altruistic donors who do not have any paired patients. This allows us to also perform kidney exchange along paths that start from some altruistic donor. Unfortunately, the computational task is NP-complete. To overcome this computational barrier, an important line of research focuses on designing faster algorithms, both exact and using the framework of parameterized complexity.\n The standard parameter for the kidney exchange problem is the number $t$ of patients that receive a healthy kidney. The current fastest known deterministic FPT algorithm for this problem, parameterized by $t$, is $O^\\star\\left(14^t\\right)$. In this work, we improve this by presenting a deterministic FPT algorithm that runs in time $O^\\star\\left((4e)^t\\right)\\approx O^\\star\\left(10.88^t\\right)$. This problem is also known to be W[1]-hard parameterized by the treewidth of the underlying undirected graph. A natural question here is whether the kidney exchange problem admits an FPT algorithm parameterized by the pathwidth of the underlying undirected graph. We answer this negatively in this paper by proving that this problem is W[1]-hard parameterized by the pathwidth of the underlying undirected graph. We also present some parameterized intractability results improving the current understanding of the problem under the framework of parameterized complexity.", "date": "2025-12-30", "categories": []} {"id": "2512.24251", "title": "Deep Reinforcement Learning for Solving the Fleet Size and Mix Vehicle Routing Problem", "abstract": "The Fleet Size and Mix Vehicle Routing Problem (FSMVRP) is a prominent variant of the Vehicle Routing Problem (VRP), extensively studied in operations research and computational science. FSMVRP requires simultaneous decisions on fleet composition and routing, making it highly applicable to real-world scenarios such as short-term vehicle rental and on-demand logistics. However, these requirements also increase the complexity of FSMVRP, posing significant challenges, particularly in large-scale and time-constrained environments. In this paper, we propose a deep reinforcement learning (DRL)-based approach for solving FSMVRP, capable of generating near-optimal solutions within a few seconds. Specifically, we formulate the problem as a Markov Decision Process (MDP) and develop a novel policy network, termed FRIPN, that seamlessly integrates fleet composition and routing decisions. Our method incorporates specialized input embeddings designed for distinctdecision objectives, including a remaining graph embedding to facilitate effective vehicle employment decisions. Comprehensive experiments are conducted on both randomly generated instances and benchmark datasets. The experimental results demonstrate that our method exhibits notable advantages in terms of computational efficiency and scalability, particularly in large-scale and time-constrained scenarios. These strengths highlight the potential of our approach for practical applications and provide valuable inspiration for extending DRL-based techniques to other variants of VRP.", "date": "2025-12-30", "categories": []} {"id": "2512.22171", "title": "Solving Multi-Agent Multi-Goal Path Finding Problems in Polynomial Time", "abstract": "In this paper, we plan missions for a fleet of agents in undirected graphs, such as grids, with multiple goals. In contrast to regular multi-agent path-finding, the solver finds and updates the assignment of goals to the agents on its own. In the continuous case for a point agent with motions in the Euclidean plane, the problem can be solved arbitrarily close to optimal. For discrete variants that incur node and edge conflicts, we show that it can be solved in polynomial time, which is unexpected, since traditional vehicle routing on general graphs is NP-hard. We implement a corresponding planner that finds conflict-free optimized routes for the agents. Global assignment strategies greatly reduce the number of conflicts, with the remaining ones resolved by elaborating on the concept of ants-on-the-stick, by solving local assignment problems, by interleaving agent paths, and by kicking agents that have already arrived out of their destinations", "date": "2025-12-17", "categories": []} {"id": "2512.23150", "title": "Constraint programming model and biased random-key genetic algorithm for the single-machine coupled task scheduling problem with exact delays to minimize the makespan", "abstract": "We consider the strongly NP-hard single-machine coupled task scheduling problem with exact delays to minimize the makespan. In this problem, a set of jobs has to be scheduled, each composed of two tasks interspersed by an exact delay. Given that no preemption is allowed, the goal consists of minimizing the completion time of the last scheduled task. We model the problem using constraint programming (CP) and propose a biased random-key genetic algorithm (BRKGA). Our CP model applies well-established global constraints. Our BRKGA combines some successful components in the literature: an initial solution generator, periodical restarts and shakes, and a local search algorithm. Furthermore, the BRKGA's decoder is focused on efficiency rather than optimality, which accelerates the solution space exploration. Computational experiments on a benchmark set containing instances with up to 100 jobs (200 tasks) indicate that the proposed BRKGA can efficiently explore the problem solution space, providing high-quality approximate solutions within low computational times. It can also provide better solutions than the CP model under the same computational settings, i.e., three minutes of time limit and a single thread. The CP model, when offered a longer running time of 3600 seconds and multiple threads, significantly improved the results, reaching the current best-known solution for 90.56% of these instances. Finally, our experiments highlight the importance of the shake and local search components in the BRKGA, whose combination significantly improves the results of a standard BRKGA.", "date": "2025-12-29", "categories": []} {"id": "2512.21654", "title": "Structural Induced Exploration for Balanced and Scalable Multi-Robot Path Planning", "abstract": "Multi-robot path planning is a fundamental yet challenging problem due to its combinatorial complexity and the need to balance global efficiency with fair task allocation among robots. Traditional swarm intelligence methods, although effective on small instances, often converge prematurely and struggle to scale to complex environments. In this work, we present a structure-induced exploration framework that integrates structural priors into the search process of the ant colony optimization (ACO). The approach leverages the spatial distribution of the task to induce a structural prior at initialization, thereby constraining the search space. The pheromone update rule is then designed to emphasize structurally meaningful connections and incorporates a load-aware objective to reconcile the total travel distance with individual robot workload. An explicit overlap suppression strategy further ensures that tasks remain distinct and balanced across the team. The proposed framework was validated on diverse benchmark scenarios covering a wide range of instance sizes and robot team configurations. The results demonstrate consistent improvements in route compactness, stability, and workload distribution compared to representative metaheuristic baselines. Beyond performance gains, the method also provides a scalable and interpretable framework that can be readily applied to logistics, surveillance, and search-and-rescue applications where reliable large-scale coordination is essential.", "date": "2025-12-25", "categories": []} {"id": "2512.21657", "title": "Near-Optimal Coalition Structures in Polynomial Time", "abstract": "We study the classical coalition structure generation (CSG) problem and compare the anytime behavior of three algorithmic paradigms: dynamic programming (DP), MILP branch-and-bound, and sparse relaxations based on greedy or $l_1$-type methods. Under a simple random \"sparse synergy\" model for coalition values, we prove that sparse relaxations recover coalition structures whose welfare is arbitrarily close to optimal in polynomial time with high probability. In contrast, broad classes of DP and MILP algorithms require exponential time before attaining comparable solution quality. This establishes a rigorous probabilistic anytime separation in favor of sparse relaxations, even though exact methods remain ultimately optimal.", "date": "2025-12-25", "categories": []} {"id": "2512.20915", "title": "Towards a General Framework for Predicting and Explaining the Hardness of Graph-based Combinatorial Optimization Problems using Machine Learning and Association Rule Mining", "abstract": "This study introduces GCO-HPIF, a general machine-learning-based framework to predict and explain the computational hardness of combinatorial optimization problems that can be represented on graphs. The framework consists of two stages. In the first stage, a dataset is created comprising problem-agnostic graph features and hardness classifications of problem instances. Machine-learning-based classification algorithms are trained to map graph features to hardness categories. In the second stage, the framework explains the predictions using an association rule mining algorithm. Additionally, machine-learning-based regression models are trained to predict algorithmic computation times. The GCO-HPIF framework was applied to a dataset of 3287 maximum clique problem instances compiled from the COLLAB, IMDB, and TWITTER graph datasets using five state-of-the-art algorithms, namely three exact branch-and-bound-based algorithms (Gurobi, CliSAT, and MOMC) and two graph-neural-network-based algorithms (EGN and HGS). The framework demonstrated excellent performance in predicting instance hardness, achieving a weighted F1 score of 0.9921, a minority-class F1 score of 0.878, and an ROC-AUC score of 0.9083 using only three graph features. The best association rule found by the FP-Growth algorithm for explaining the hardness predictions had a support of 0.8829 for hard instances and an overall accuracy of 87.64 percent, underscoring the framework's usefulness for both prediction and explanation. Furthermore, the best-performing regression model for predicting computation times achieved a percentage RMSE of 5.12 and an R2 value of 0.991.", "date": "2025-12-24", "categories": []} {"id": "2512.19882", "title": "A Branch-and-Price Algorithm for Fast and Equitable Last-Mile Relief Aid Distribution", "abstract": "The distribution of relief supplies to shelters is a critical aspect of post-disaster humanitarian logistics. In major disasters, prepositioned supplies often fall short of meeting all demands. We address the problem of planning vehicle routes from a distribution center to shelters while allocating limited relief supplies. To balance efficiency and equity, we formulate a bi-objective problem: minimizing a Gini-index-based measure of inequity in unsatisfied demand for fair distribution and minimizing total travel time for timely delivery. We propose a Mixed Integer Programming (MIP) model and use the $\\epsilon$-constraint method to handle the bi-objective nature. By deriving mathematical properties of the optimal solution, we introduce valid inequalities and design an algorithm for optimal delivery allocations given feasible vehicle routes. A branch-and-price (B&P) algorithm is developed to solve the problem efficiently. Computational tests on realistic datasets from a past earthquake in Van, Turkey, and predicted data for Istanbul's Kartal region show that the B&P algorithm significantly outperforms commercial MIP solvers. Our bi-objective approach reduces aid distribution inequity by 34% without compromising efficiency. Results indicate that when time constraints are very loose or tight, lexicographic optimization prioritizing demand coverage over fairness is effective. For moderately restrictive time constraints, a balanced approach is essential to avoid inequitable outcomes.", "date": "2025-12-22", "categories": []} {"id": "2407.20777", "title": "Feature-Guided Metaheuristic with Diversity Management for Solving the Capacitated Vehicle Routing Problem", "abstract": "We propose a feature-based guidance mechanism to enhance metaheuristic algorithms for solving the Capacitated Vehicle Routing Problem (CVRP). This mechanism leverages an Explainable AI (XAI) model to identify features that correlate with high-quality solutions. These insights are used to guide the search process by promoting solution diversity and avoiding premature convergence. The guidance mechanism is first integrated into a custom metaheuristic algorithm, which combines neighborhood search with a novel hybrid of the split algorithm and path relinking. Experiments on benchmark instances with up to $30,000$ customer nodes demonstrate that the guidance significantly improves the performance of this baseline algorithm. Furthermore, we validate the generalizability of the guidance approach by integrating it into a state-of-the-art metaheuristic, where it again yields statistically significant performance gains. These results confirm that the proposed mechanism is both scalable and transferable across algorithmic frameworks.", "date": "2024-07-30", "categories": []} {"id": "2512.11469", "title": "Three methods, one problem: Classical and AI approaches to no-three-in-line", "abstract": "The No-Three-In-Line problem asks for the maximum number of points that can be placed on an n by n grid with no three collinear, representing a famous problem in combinatorial geometry. While classical methods like Integer Linear Programming (ILP) guarantee optimal solutions, they face exponential scaling with grid size, and recent advances in machine learning offer promising alternatives for pattern-based approximation. This paper presents the first systematic comparison of classical optimization and AI approaches to this problem, evaluating their performance against traditional algorithms. We apply PatternBoost transformer learning and reinforcement learning (PPO) to this problem for the first time, comparing them against ILP. ILP achieves provably optimal solutions up to 19 by 19 grids, while PatternBoost matches optimal performance up to 14 by 14 grids with 96% test loss reduction. PPO achieves perfect solutions on 10 by 10 grids but fails at 11 by 11 grids, where constraint violations prevent valid configurations. These results demonstrate that classical optimization remains essential for exact solutions while AI methods offer competitive performance on smaller instances, with hybrid approaches presenting the most promising direction for scaling to larger problem sizes.", "date": "2025-12-12", "categories": []} {"id": "2512.18618", "title": "Assignment-Routing Optimization: Solvers for Problems Under Constraints", "abstract": "We study the Joint Routing-Assignment (JRA) problem in which items must be assigned one-to-one to placeholders while simultaneously determining a Hamiltonian cycle visiting all nodes exactly once. Extending previous exact MIP solvers with Gurobi and cutting-plane subtour elimination, we develop a solver tailored for practical packaging-planning scenarios with richer constraints.These include multiple placeholder options, time-frame restrictions, and multi-class item packaging. Experiments on 46 mobile manipulation datasets demonstrate that the proposed MIP approach achieves global optima with stable and low computation times, significantly outperforming the shaking-based exact solver by up to an orders of magnitude. Compared to greedy baselines, the MIP solutions achieve consistent optimal distances with an average deviation of 14% for simple heuristics, confirming both efficiency and solution quality. The results highlight the practical applicability of MIP-based JRA optimization for robotic packaging, motion planning, and complex logistics .", "date": "2025-12-21", "categories": []} {"id": "2512.18633", "title": "ARC: Leveraging Compositional Representations for Cross-Problem Learning on VRPs", "abstract": "Vehicle Routing Problems (VRPs) with diverse real-world attributes have driven recent interest in cross-problem learning approaches that efficiently generalize across problem variants. We propose ARC (Attribute Representation via Compositional Learning), a cross-problem learning framework that learns disentangled attribute representations by decomposing them into two complementary components: an Intrinsic Attribute Embedding (IAE) for invariant attribute semantics and a Contextual Interaction Embedding (CIE) for attribute-combination effects. This disentanglement is achieved by enforcing analogical consistency in the embedding space to ensure the semantic transformation of adding an attribute (e.g., a length constraint) remains invariant across different problem contexts. This enables our model to reuse invariant semantics across trained variants and construct representations for unseen combinations. ARC achieves state-of-the-art performance across in-distribution, zero-shot generalization, few-shot adaptation, and real-world benchmarks.", "date": "2025-12-21", "categories": []} {"id": "2512.19527", "title": "Deep Learning for Unrelated-Machines Scheduling: Handling Variable Dimensions", "abstract": "Deep learning has been effectively applied to many discrete optimization problems. However, learning-based scheduling on unrelated parallel machines remains particularly difficult to design. Not only do the numbers of jobs and machines vary, but each job-machine pair has a unique processing time, dynamically altering feature dimensions. We propose a novel approach with a neural network tailored for offline deterministic scheduling of arbitrary sizes on unrelated machines. The goal is to minimize a complex objective function that includes the makespan and the weighted tardiness of jobs and machines. Unlike existing online approaches, which process jobs sequentially, our method generates a complete schedule considering the entire input at once. The key contribution of this work lies in the sophisticated architecture of our model. By leveraging various NLP-inspired architectures, it effectively processes any number of jobs and machines with varying feature dimensions imposed by unrelated processing times. Our approach enables supervised training on small problem instances while demonstrating strong generalization to much larger scheduling environments. Trained and tested on instances with 8 jobs and 4 machines, costs were only 2.51% above optimal. Across all tested configurations of up to 100 jobs and 10 machines, our network consistently outperformed an advanced dispatching rule, which incurred 22.22% higher costs on average. As our method allows fast retraining with simulated data and adaptation to various scheduling conditions, we believe it has the potential to become a standard approach for learning-based scheduling on unrelated machines and similar problem environments.", "date": "2025-12-22", "categories": []} {"id": "2502.00277", "title": "Regularized Langevin Dynamics for Combinatorial Optimization", "abstract": "This work proposes a simple yet effective sampling framework for combinatorial optimization (CO). Our method builds on discrete Langevin dynamics (LD), an efficient gradient-guided generative paradigm. However, we observe that directly applying LD often leads to limited exploration. To overcome this limitation, we propose the Regularized Langevin Dynamics (RLD), which enforces an expected distance between the sampled and current solutions, effectively avoiding local minima. We develop two CO solvers on top of RLD, one based on simulated annealing (SA), and the other one based on neural network (NN). Empirical results on three classic CO problems demonstrate that both of our methods can achieve comparable or better performance against the previous state-of-the-art (SOTA) SA- and NN-based solvers. In particular, our SA algorithm reduces the runtime of the previous SOTA SA method by up to 80\\%, while achieving equal or superior performance. In summary, RLD offers a promising framework for enhancing both traditional heuristics and NN models to solve CO problems. Our code is available at https://github.com/Shengyu-Feng/RLD4CO.", "date": "2025-02-01", "categories": []} {"id": "2512.17412", "title": "Optimisation of Aircraft Maintenance Schedules", "abstract": "We present an aircraft maintenance scheduling problem, which requires suitably qualified staff to be assigned to maintenance tasks on each aircraft. The tasks on each aircraft must be completed within a given turn around window so that the aircraft may resume revenue earning service. This paper presents an initial study based on the application of an Evolutionary Algorithm to the problem. Evolutionary Algorithms evolve a solution to a problem by evaluating many possible solutions, focusing the search on those solutions that are of a higher quality, as defined by a fitness function. In this paper, we benchmark the algorithm on 60 generated problem instances to demonstrate the underlying representation and associated genetic operators.", "date": "2025-12-19", "categories": []} {"id": "2512.03419", "title": "Comparative algorithm performance evaluation and prediction for the maximum clique problem using instance space analysis", "abstract": "The maximum clique problem, a well-known graph-based combinatorial optimization problem, has been addressed through various algorithmic approaches, though systematic analyses of the problem instances remain sparse. This study employs the instance space analysis (ISA) methodology to systematically analyze the instance space of this problem and assess & predict the performance of state-of-the-art (SOTA) algorithms, including exact, heuristic, and graph neural network (GNN)-based methods. A dataset was compiled using graph instances from TWITTER, COLLAB and IMDB-BINARY benchmarks commonly used in graph machine learning research. A set of 33 generic and 2 problem-specific polynomial-time-computable graph-based features, including several spectral properties, was employed for the ISA. A composite performance measure incorporating both solution quality and algorithm runtime was utilized. The comparative analysis demonstrated that the exact algorithm Mixed Order Maximum Clique (MOMC) exhibited superior performance across approximately 74.7% of the instance space constituted by the compiled dataset. Gurobi & CliSAT accounted for superior performance in 13.8% and 11% of the instance space, respectively. The ISA-based algorithm performance prediction model run on 34 challenging test instances compiled from the BHOSLIB and DIMACS datasets yielded top-1 and top-2 best performing algorithm prediction accuracies of 88% and 97%, respectively.", "date": "2025-12-03", "categories": []} {"id": "2512.15800", "title": "Edge-wise Topological Divergence Gaps: Guiding Search in Combinatorial Optimization", "abstract": "We introduce a topological feedback mechanism for the Travelling Salesman Problem (TSP) by analyzing the divergence between a tour and the minimum spanning tree (MST). Our key contribution is a canonical decomposition theorem that expresses the tour-MST gap as edge-wise topology-divergence gaps from the RTD-Lite barcode. Based on this, we develop a topological guidance for 2-opt and 3-opt heuristics that increases their performance. We carry out experiments with fine-optimization of tours obtained from heatmap-based methods, TSPLIB, and random instances. Experiments demonstrate the topology-guided optimization results in better performance and faster convergence in many cases.", "date": "2025-12-16", "categories": []} {"id": "2511.17435", "title": "Multi-Agent Pointer Transformer: Seq-to-Seq Reinforcement Learning for Multi-Vehicle Dynamic Pickup-Delivery Problems", "abstract": "This paper addresses the cooperative Multi-Vehicle Dynamic Pickup and Delivery Problem with Stochastic Requests (MVDPDPSR) and proposes an end-to-end centralized decision-making framework based on sequence-to-sequence, named Multi-Agent Pointer Transformer (MAPT). MVDPDPSR is an extension of the vehicle routing problem and a spatio-temporal system optimization problem, widely applied in scenarios such as on-demand delivery. Classical operations research methods face bottlenecks in computational complexity and time efficiency when handling large-scale dynamic problems. Although existing reinforcement learning methods have achieved some progress, they still encounter several challenges: 1) Independent decoding across multiple vehicles fails to model joint action distributions; 2) The feature extraction network struggles to capture inter-entity relationships; 3) The joint action space is exponentially large. To address these issues, we designed the MAPT framework, which employs a Transformer Encoder to extract entity representations, combines a Transformer Decoder with a Pointer Network to generate joint action sequences in an AutoRegressive manner, and introduces a Relation-Aware Attention module to capture inter-entity relationships. Additionally, we guide the model's decision-making using informative priors to facilitate effective exploration. Experiments on 8 datasets demonstrate that MAPT significantly outperforms existing baseline methods in terms of performance and exhibits substantial computational time advantages compared to classical operations research methods.", "date": "2025-11-21", "categories": []} {"id": "2512.15210", "title": "Label-consistent clustering for evolving data", "abstract": "Data analysis often involves an iterative process, where solutions must be continuously refined in response to new data. Typically, as new data becomes available, an existing solution must be updated to incorporate the latest information. In addition to seeking a high-quality solution for the task at hand, it is also crucial to ensure consistency by minimizing drastic changes from previous solutions. Applying this approach across many iterations, ensures that the solution evolves gradually and smoothly.\n In this paper, we study the above problem in the context of clustering, specifically focusing on the $k$-center problem. More precisely, we study the following problem: Given a set of points $X$, parameters $k$ and $b$, and a prior clustering solution $H$ for $X$, our goal is to compute a new solution $C$ for $X$, consisting of $k$ centers, which minimizes the clustering cost while introducing at most $b$ changes from $H$. We refer to this problem as label-consistent $k$-center, and we propose two constant-factor approximation algorithms for it. We complement our theoretical findings with an experimental evaluation demonstrating the effectiveness of our methods on real-world datasets.", "date": "2025-12-17", "categories": []} {"id": "2512.11864", "title": "Solving Parallel Machine Scheduling With Precedences and Cumulative Resource Constraints With Calendars", "abstract": "The task of finding efficient production schedules for parallel machines is a challenge that arises in most industrial manufacturing domains. There is a large potential to minimize production costs through automated scheduling techniques, due to the large-scale requirements of modern factories. In the past, solution approaches have been studied for many machine scheduling variations, where even basic variants have been shown to be NP-hard. However, in today's real-life production environments, additional complex precedence constraints and resource restrictions with calendars arise that must be fulfilled. These additional constraints cannot be tackled efficiently by existing solution techniques. Thus, there is a strong need to develop and analyze automated methods that can solve such real-life parallel machine scheduling scenarios. In this work, we introduce a novel variant of parallel machine scheduling with job precedences and calendar-based cumulative resource constraints that arises in real-life industrial use cases. A constraint modeling approach is proposed as an exact solution method for small scheduling scenarios together with state-of-the-art constraint-solving technology. Further, we propose a construction heuristic as well as a tailored metaheuristic using local search to efficiently tackle large-scale problem instances. This metaheuristic approach has been deployed and is currently being used in an industrial setting.", "date": "2025-12-05", "categories": []} {"id": "2510.21525", "title": "A Unified Model for Multi-Task Drone Routing in Post-Disaster Road Assessment", "abstract": "Post-disaster road assessment (PDRA) is essential for emergency response, enabling rapid evaluation of infrastructure conditions and efficient allocation of resources. Although drones provide a flexible and effective tool for PDRA, routing them in large-scale networks remains challenging. Exact and heuristic optimization methods scale poorly and demand domain expertise, while existing deep reinforcement learning (DRL) approaches adopt a single-task paradigm, requiring separate models for each problem variant and lacking adaptability to evolving operational needs. This study proposes a unified model (UM) for drone routing that simultaneously addresses eight PDRA variants. By training a single neural network across multiple problem configurations, UM captures shared structural knowledge while adapting to variant-specific constraints through a modern transformer encoder-decoder architecture. A lightweight adapter mechanism further enables efficient finetuning to unseen attributes without retraining, enhancing deployment flexibility in dynamic disaster scenarios. Extensive experiments demonstrate that the UM reduces training time and parameters by a factor of eight compared with training separate models, while consistently outperforming single-task DRL methods by 6-14%, heuristic algorithms by 22-42%, and commercial solvers by 24-82% in terms of solution quality (total collected information value). The model achieves rapid solutions (1-10 seconds) across networks of up to 1,000 nodes, with robustness confirmed through sensitivity analyses. Moreover, finetuning experiments show that unseen attributes can be effectively incorporated with minimal cost while retaining high solution quality. The source code for UM is publicly available at https://github.com/PJ-HTU/UM_PDRA.", "date": "2025-10-24", "categories": []} {"id": "2512.11169", "title": "CORL: Reinforcement Learning of MILP Policies Solved via Branch and Bound", "abstract": "Combinatorial sequential decision making problems are typically modeled as mixed integer linear programs (MILPs) and solved via branch and bound (B&B) algorithms. The inherent difficulty of modeling MILPs that accurately represent stochastic real world problems leads to suboptimal performance in the real world. Recently, machine learning methods have been applied to build MILP models for decision quality rather than how accurately they model the real world problem. However, these approaches typically rely on supervised learning, assume access to true optimal decisions, and use surrogates for the MILP gradients. In this work, we introduce a proof of concept CORL framework that end to end fine tunes an MILP scheme using reinforcement learning (RL) on real world data to maximize its operational performance. We enable this by casting an MILP solved by B&B as a differentiable stochastic policy compatible with RL. We validate the CORL method in a simple illustrative combinatorial sequential decision making example.", "date": "2025-12-11", "categories": []} {"id": "2512.11187", "title": "Deep Learning--Accelerated Multi-Start Large Neighborhood Search for Real-time Freight Bundling", "abstract": "Online Freight Exchange Systems (OFEX) play a crucial role in modern freight logistics by facilitating real-time matching between shippers and carrier. However, efficient combinatorial bundling of transporation jobs remains a bottleneck. We model the OFEX combinatorial bundling problem as a multi-commodity one-to-one pickup-and-delivery selective traveling salesperson problem (m1-PDSTSP), which optimizes revenue-driven freight bundling under capacity, precedence, and route-length constraints. The key challenge is to couple combinatorial bundle selection with pickup-and-delivery routing under sub-second latency. We propose a learning--accelerated hybrid search pipeline that pairs a Transformer Neural Network-based constructive policy with an innovative Multi-Start Large Neighborhood Search (MSLNS) metaheuristic within a rolling-horizon scheme in which the platform repeatedly freezes the current marketplace into a static snapshot and solves it under a short time budget. This pairing leverages the low-latency, high-quality inference of the learning-based constructor alongside the robustness of improvement search; the multi-start design and plausible seeds help LNS to explore the solution space more efficiently. Across benchmarks, our method outperforms state-of-the-art neural combinatorial optimization and metaheuristic baselines in solution quality with comparable time, achieving an optimality gap of less than 2\\% in total revenue relative to the best available exact baseline method. To our knowledge, this is the first work to establish that a Deep Neural Network-based constructor can reliably provide high-quality seeds for (multi-start) improvement heuristics, with applicability beyond the \\textit{m1-PDSTSP} to a broad class of selective traveling salesperson problems and pickup and delivery problems.", "date": "2025-12-12", "categories": []} {"id": "2512.10813", "title": "Quantum Approaches to Urban Logistics: From Core QAOA to Clustered Scalability", "abstract": "The Traveling Salesman Problem (TSP) is a fundamental challenge in combinatorial optimization, widely applied in logistics and transportation. As the size of TSP instances grows, traditional algorithms often struggle to produce high-quality solutions within reasonable timeframes. This study investigates the potential of the Quantum Approximate Optimization Algorithm (QAOA), a hybrid quantum-classical method, to solve TSP under realistic constraints. We adopt a QUBO-based formulation of TSP that integrates real-world logistical constraints reflecting operational conditions, such as vehicle capacity, road accessibility, and time windows, while ensuring compatibility with the limitations of current quantum hardware. Our experiments are conducted in a simulated environment using high-performance computing (HPC) resources to assess QAOA's performance across different problem sizes and quantum circuit depths. In order to improve scalability, we propose clustering QAOA (Cl-QAOA), a hybrid approach combining classical machine learning with QAOA. This method decomposes large TSP instances into smaller sub-problems, making quantum optimization feasible even on devices with a limited number of qubits. The results offer a comprehensive evaluation of QAOA's strengths and limitations in solving constrained TSP scenarios. This study advances quantum optimization and lays groundwork for future large-scale applications.", "date": "2025-12-11", "categories": []} {"id": "2512.09318", "title": "Simultaneous Genetic Evolution of Neural Networks for Optimal SFC Embedding", "abstract": "The reliance of organisations on computer networks is enabled by network programmability, which is typically achieved through Service Function Chaining. These chains virtualise network functions, link them, and programmatically embed them on networking infrastructure. Optimal embedding of Service Function Chains is an NP-hard problem, with three sub-problems, chain composition, virtual network function embedding, and link embedding, that have to be optimised simultaneously, rather than sequentially, for optimal results. Genetic Algorithms have been employed for this, but existing approaches either do not optimise all three sub-problems or do not optimise all three sub-problems simultaneously. We propose a Genetic Algorithm-based approach called GENESIS, which evolves three sine-function-activated Neural Networks, and funnels their output to a Gaussian distribution and an A* algorithm to optimise all three sub-problems simultaneously. We evaluate GENESIS on an emulator across 48 different data centre scenarios and compare its performance to two state-of-the-art Genetic Algorithms and one greedy algorithm. GENESIS produces an optimal solution for 100% of the scenarios, whereas the second-best method optimises only 71% of the scenarios. Moreover, GENESIS is the fastest among all Genetic Algorithms, averaging 15.84 minutes, compared to an average of 38.62 minutes for the second-best Genetic Algorithm.", "date": "2025-12-10", "categories": []} {"id": "2305.17705", "title": "Towards Autonomous and Safe Last-mile Deliveries with AI-augmented Self-driving Delivery Robots", "abstract": "In addition to its crucial impact on customer satisfaction, last-mile delivery (LMD) is notorious for being the most time-consuming and costly stage of the shipping process. Pressing environmental concerns combined with the recent surge of e-commerce sales have sparked renewed interest in automation and electrification of last-mile logistics. To address the hurdles faced by existing robotic couriers, this paper introduces a customer-centric and safety-conscious LMD system for small urban communities based on AI-assisted autonomous delivery robots. The presented framework enables end-to-end automation and optimization of the logistic process while catering for real-world imposed operational uncertainties, clients' preferred time schedules, and safety of pedestrians. To this end, the integrated optimization component is modeled as a robust variant of the Cumulative Capacitated Vehicle Routing Problem with Time Windows, where routes are constructed under uncertain travel times with an objective to minimize the total latency of deliveries (i.e., the overall waiting time of customers, which can negatively affect their satisfaction). We demonstrate the proposed LMD system's utility through real-world trials in a university campus with a single robotic courier. Implementation aspects as well as the findings and practical insights gained from the deployment are discussed in detail. Lastly, we round up the contributions with numerical simulations to investigate the scalability of the developed mathematical formulation with respect to the number of robotic vehicles and customers.", "date": "2023-05-28", "categories": []} {"id": "2508.03929", "title": "MOTIF: Multi-strategy Optimization via Turn-based Interactive Framework", "abstract": "Designing effective algorithmic components remains a fundamental obstacle in tackling NP-hard combinatorial optimization problems (COPs), where solvers often rely on carefully hand-crafted strategies. Despite recent advances in using large language models (LLMs) to synthesize high-quality components, most approaches restrict the search to a single element - commonly a heuristic scoring function - thus missing broader opportunities for innovation. In this paper, we introduce a broader formulation of solver design as a multi-strategy optimization problem, which seeks to jointly improve a set of interdependent components under a unified objective. To address this, we propose Multi-strategy Optimization via Turn-based Interactive Framework (MOTIF) - a novel framework based on Monte Carlo Tree Search that facilitates turn-based optimization between two LLM agents. At each turn, an agent improves one component by leveraging the history of both its own and its opponent's prior updates, promoting both competitive pressure and emergent cooperation. This structured interaction broadens the search landscape and encourages the discovery of diverse, high-performing solutions. Experiments across multiple COP domains show that MOTIF consistently outperforms state-of-the-art methods, highlighting the promise of turn-based, multi-agent prompting for fully automated solver design.", "date": "2025-08-05", "categories": []} {"id": "2512.06351", "title": "LLM-Upgraded Graph Reinforcement Learning for Carbon-Aware Job Scheduling in Smart Manufacturing", "abstract": "This paper presents \\textsc{Luca}, a \\underline{l}arge language model (LLM)-\\underline{u}pgraded graph reinforcement learning framework for \\underline{c}arbon-\\underline{a}ware flexible job shop scheduling. \\textsc{Luca} addresses the challenges of dynamic and sustainable scheduling in smart manufacturing systems by integrating a graph neural network and an LLM, guided by a carefully designed in-house prompting strategy, to produce a fused embedding that captures both structural characteristics and contextual semantics of the latest scheduling state. This expressive embedding is then processed by a deep reinforcement learning policy network, which generates real-time scheduling decisions optimized for both makespan and carbon emission objectives. To support sustainability goals, \\textsc{Luca} incorporates a dual-objective reward function that encourages both energy efficiency and scheduling timeliness. Experimental results on both synthetic and public datasets demonstrate that \\textsc{Luca} consistently outperforms comparison algorithms. For instance, on the synthetic dataset, it achieves an average of 4.1\\% and up to 12.2\\% lower makespan compared to the best-performing comparison algorithm while maintaining the same emission level. On public datasets, additional gains are observed for both makespan and emission. These results demonstrate that \\textsc{Luca} is effective and practical for carbon-aware scheduling in smart manufacturing.", "date": "2025-12-06", "categories": []} {"id": "2512.05619", "title": "Enhancing Local Search for MaxSAT with Deep Differentiation Clause Weighting", "abstract": "Partial Maximum Satisfiability (PMS) and Weighted Partial Maximum Satisfiability (WPMS) generalize Maximum Satisfiability (MaxSAT), with broad real-world applications. Recent advances in Stochastic Local Search (SLS) algorithms for solving (W)PMS have mainly focused on designing clause weighting schemes. However, existing methods often fail to adequately distinguish between PMS and WPMS, typically employing uniform update strategies for clause weights and overlooking critical structural differences between the two problem types. In this work, we present a novel clause weighting scheme that, for the first time, updates the clause weights of PMS and WPMS instances according to distinct conditions. This scheme also introduces a new initialization method, which better accommodates the unique characteristics of both instance types. Furthermore, we propose a decimation method that prioritizes satisfying unit and hard clauses, effectively complementing our proposed clause weighting scheme. Building on these methods, we develop a new SLS solver for (W)PMS named DeepDist. Experimental results on benchmarks from the anytime tracks of recent MaxSAT Evaluations show that DeepDist outperforms state-of-the-art SLS solvers. Notably, a hybrid solver combining DeepDist with TT-Open-WBO-Inc surpasses the performance of the MaxSAT Evaluation 2024 winners, SPB-MaxSAT-c-Band and SPB-MaxSAT-c-FPS, highlighting the effectiveness of our approach. The code is available at https://github.com/jmhmaxsat/DeepDist", "date": "2025-12-05", "categories": []} {"id": "2512.05946", "title": "Variational Quantum Rainbow Deep Q-Network for Optimizing Resource Allocation Problem", "abstract": "Resource allocation remains NP-hard due to combinatorial complexity. While deep reinforcement learning (DRL) methods, such as the Rainbow Deep Q-Network (DQN), improve scalability through prioritized replay and distributional heads, classical function approximators limit their representational power. We introduce Variational Quantum Rainbow DQN (VQR-DQN), which integrates ring-topology variational quantum circuits with Rainbow DQN to leverage quantum superposition and entanglement. We frame the human resource allocation problem (HRAP) as a Markov decision process (MDP) with combinatorial action spaces based on officer capabilities, event schedules, and transition times. On four HRAP benchmarks, VQR-DQN achieves 26.8% normalized makespan reduction versus random baselines and outperforms Double DQN and classical Rainbow DQN by 4.9-13.4%. These gains align with theoretical connections between circuit expressibility, entanglement, and policy quality, demonstrating the potential of quantum-enhanced DRL for large-scale resource allocation. Our implementation is available at: https://github.com/Analytics-Everywhere-Lab/qtrl/.", "date": "2025-12-05", "categories": []} {"id": "2412.16181", "title": "Minimum Weighted Feedback Arc Sets for Ranking from Pairwise Comparisons", "abstract": "The Minimum Weighted Feedback Arc Set (MWFAS) problem is closely related to the task of deriving a global ranking from pairwise comparisons. Recent work by He et al. (ICML 2022) advanced the state of the art on ranking benchmarks using learning based methods, but did not examine the underlying connection to MWFAS. In this paper, we investigate this relationship and introduce efficient combinatorial algorithms for solving MWFAS as a means of addressing the ranking problem. Our experimental results show that these simple, learning free methods achieve substantially faster runtimes than recent learning based approaches, while also delivering competitive, and in many cases superior, ranking accuracy. These findings suggest that lightweight combinatorial techniques offer a scalable and effective alternative to deep learning for large scale ranking tasks.", "date": "2024-12-10", "categories": []} {"id": "2512.03762", "title": "RoCo: Role-Based LLMs Collaboration for Automatic Heuristic Design", "abstract": "Automatic Heuristic Design (AHD) has gained traction as a promising solution for solving combinatorial optimization problems (COPs). Large Language Models (LLMs) have emerged and become a promising approach to achieving AHD, but current LLM-based AHD research often only considers a single role. This paper proposes RoCo, a novel Multi-Agent Role-Based System, to enhance the diversity and quality of AHD through multi-role collaboration. RoCo coordinates four specialized LLM-guided agents-explorer, exploiter, critic, and integrator-to collaboratively generate high-quality heuristics. The explorer promotes long-term potential through creative, diversity-driven thinking, while the exploiter focuses on short-term improvements via conservative, efficiency-oriented refinements. The critic evaluates the effectiveness of each evolution step and provides targeted feedback and reflection. The integrator synthesizes proposals from the explorer and exploiter, balancing innovation and exploitation to drive overall progress. These agents interact in a structured multi-round process involving feedback, refinement, and elite mutations guided by both short-term and accumulated long-term reflections. We evaluate RoCo on five different COPs under both white-box and black-box settings. Experimental results demonstrate that RoCo achieves superior performance, consistently generating competitive heuristics that outperform existing methods including ReEvo and HSEvo, both in white-box and black-box scenarios. This role-based collaborative paradigm establishes a new standard for robust and high-performing AHD.", "date": "2025-12-03", "categories": []} {"id": "2512.04139", "title": "Solving N-Queen Problem using Las Vegas Algorithm with State Pruning", "abstract": "The N-Queens problem, placing all N queens in a N x N chessboard where none attack the other, is a classic problem for constraint satisfaction algorithms. While complete methods like backtracking guarantee a solution, their exponential time complexity makes them impractical for large-scale instances thus, stochastic approaches, such as Las Vegas algorithm, are preferred. While it offers faster approximate solutions, it suffers from significant performance variance due to random placement of queens on the board. This research introduces a hybrid algorithm built on top of the standard Las Vegas framework through iterative pruning, dynamically eliminating invalid placements during the random assignment phase, thus this method effectively reduces the search space. The analysis results that traditional backtracking scales poorly with increasing N. In contrast, the proposed technique consistently generates valid solutions more rapidly, establishing it as a superior alternative to use where a single, timely solution is preferred over completeness. Although large N causes some performance variability, the algorithm demonstrates a highly effective trade-off between computational cost and solution fidelity, making it particularly suited for resource-constrained computing environments.", "date": "2025-12-03", "categories": []} {"id": "2511.07850", "title": "GAMA: A Neural Neighborhood Search Method with Graph-aware Multi-modal Attention for Vehicle Routing Problem", "abstract": "Recent advances in neural neighborhood search methods have shown potential in tackling Vehicle Routing Problems (VRPs). However, most existing approaches rely on simplistic state representations and fuse heterogeneous information via naive concatenation, limiting their ability to capture rich structural and semantic context. To address these limitations, we propose GAMA, a neural neighborhood search method with Graph-aware Multi-modal Attention model in VRP. GAMA encodes the problem instance and its evolving solution as distinct modalities using graph neural networks, and models their intra- and inter-modal interactions through stacked self- and cross-attention layers. A gated fusion mechanism further integrates the multi-modal representations into a structured state, enabling the policy to make informed and generalizable operator selection decisions. Extensive experiments conducted across various synthetic and benchmark instances demonstrate that the proposed algorithm GAMA significantly outperforms the recent neural baselines. Further ablation studies confirm that both the multi-modal attention mechanism and the gated fusion design play a key role in achieving the observed performance gains.", "date": "2025-11-11", "categories": []} {"id": "2410.12636", "title": "Towards Arbitrary QUBO Optimization: Analysis of Classical and Quantum-Activated Feedforward Neural Networks", "abstract": "Quadratic Unconstrained Binary Optimization (QUBO) sits at the heart of many industries and academic fields such as logistics, supply chain, finance, pharmaceutical science, chemistry, IT, and energy sectors, among others. These problems typically involve optimizing a large number of binary variables, which makes finding exact solutions exponentially more difficult. Consequently, most QUBO problems are classified as NP-hard. To address this challenge, we developed a powerful feedforward neural network (FNN) optimizer for arbitrary QUBO problems. In this work, we demonstrate that the FNN optimizer can provide high-quality approximate solutions for large problems, including dense 80-variable weighted MaxCut and random QUBOs, achieving an average accuracy of over 99% in less than 1.1 seconds on an 8-core CPU. Additionally, the FNN optimizer outperformed the Gurobi optimizer by 72% on 200-variable random QUBO problems within a 100-second computation time limit, exhibiting strong potential for real-time optimization tasks. Building on this model, we explored the novel approach of integrating FNNs with a quantum annealer-based activation function to create a quantum-classical encoder-decoder (QCED) optimizer, aiming to further enhance the performance of FNNs in QUBO optimization.", "date": "2024-10-16", "categories": []} {"id": "2512.00410", "title": "Balancing Efficiency and Fairness: An Iterative Exchange Framework for Multi-UAV Cooperative Path Planning", "abstract": "Multi-UAV cooperative path planning (MUCPP) is a fundamental problem in multi-agent systems, aiming to generate collision-free trajectories for a team of unmanned aerial vehicles (UAVs) to complete distributed tasks efficiently. A key challenge lies in achieving both efficiency, by minimizing total mission cost, and fairness, by balancing the workload among UAVs to avoid overburdening individual agents. This paper presents a novel Iterative Exchange Framework for MUCPP, balancing efficiency and fairness through iterative task exchanges and path refinements. The proposed framework formulates a composite objective that combines the total mission distance and the makespan, and iteratively improves the solution via local exchanges under feasibility and safety constraints. For each UAV, collision-free trajectories are generated using A* search over a terrain-aware configuration space. Comprehensive experiments on multiple terrain datasets demonstrate that the proposed method consistently achieves superior trade-offs between total distance and makespan compared to existing baselines.", "date": "2025-11-29", "categories": []} {"id": "2512.00605", "title": "Efficient Matroid Bandit Linear Optimization Leveraging Unimodality", "abstract": "We study the combinatorial semi-bandit problem under matroid constraints. The regret achieved by recent approaches is optimal, in the sense that it matches the lower bound. Yet, time complexity remains an issue for large matroids or for matroids with costly membership oracles (e.g. online recommendation that ensures diversity). This paper sheds a new light on the matroid semi-bandit problem by exploiting its underlying unimodal structure. We demonstrate that, with negligible loss in regret, the number of iterations involving the membership oracle can be limited to \\mathcal{O}(\\log \\log T)$. This results in an overall improved time complexity of the learning process. Experiments conducted on various matroid benchmarks show (i) no loss in regret compared to state-of-the-art approaches; and (ii) reduced time complexity and number of calls to the membership oracle.", "date": "2025-11-29", "categories": []} {"id": "2512.01331", "title": "A Fast Heuristic Search Approach for Energy-Optimal Profile Routing for Electric Vehicles", "abstract": "We study the energy-optimal shortest path problem for electric vehicles (EVs) in large-scale road networks, where recuperated energy along downhill segments introduces negative energy costs. While traditional point-to-point pathfinding algorithms for EVs assume a known initial energy level, many real-world scenarios involving uncertainty in available energy require planning optimal paths for all possible initial energy levels, a task known as energy-optimal profile search. Existing solutions typically rely on specialized profile-merging procedures within a label-correcting framework that results in searching over complex profiles. In this paper, we propose a simple yet effective label-setting approach based on multi-objective A* search, which employs a novel profile dominance rule to avoid generating and handling complex profiles. We develop four variants of our method and evaluate them on real-world road networks enriched with realistic energy consumption data. Experimental results demonstrate that our energy profile A* search achieves performance comparable to energy-optimal A* with a known initial energy level.", "date": "2025-12-01", "categories": []} {"id": "2512.01518", "title": "End-to-end Deep Reinforcement Learning for Stochastic Multi-objective Optimization in C-VRPTW", "abstract": "In this work, we consider learning-based applications in routing to solve a Vehicle Routing variant characterized by stochasticity and multiple objectives. Such problems are representative of practical settings where decision-makers have to deal with uncertainty in the operational environment as well as multiple conflicting objectives due to different stakeholders. We specifically consider travel time uncertainty. We also consider two objectives, total travel time and route makespan, that jointly target operational efficiency and labor regulations on shift length, although different objectives could be incorporated. Learning-based methods offer earnest computational advantages as they can repeatedly solve problems with limited interference from the decision-maker. We specifically focus on end-to-end deep learning models that leverage the attention mechanism and multiple solution trajectories. These models have seen several successful applications in routing problems. However, since travel times are not a direct input to these models due to the large dimensions of the travel time matrix, accounting for uncertainty is a challenge, especially in the presence of multiple objectives. In turn, we propose a model that simultaneously addresses stochasticity and multi-objectivity and provide a refined training mechanism for this model through scenario clustering to reduce training time. Our results show that our model is capable of constructing a Pareto Front of good quality within acceptable run times compared to three baselines.", "date": "2025-12-01", "categories": []} {"id": "2505.03323", "title": "Unraveling the Rainbow: can value-based methods schedule?", "abstract": "In this work, we conduct an extensive empirical study of several deep reinforcement learning algorithms on two challenging combinatorial optimization problems: the job-shop and flexible job-shop scheduling problems, both fundamental challenges with multiple industrial applications. Broadly, deep reinforcement learning algorithms fall into two categories: policy-gradient and value-based. While value-based algorithms have achieved notable success in domains such as the Arcade Learning Environment, the combinatorial optimization community has predominantly favored policy-gradient algorithms, often overlooking the potential of value-based alternatives. From our results, value-based algorithms demonstrated a lower variance and a more stable convergence profile compared to policy-gradient ones. Moreover, they achieved superior cross-size and cross-distribution generalization, that is, effectively solving instances that are substantially larger or structurally distinct from those seen during training. Finally, our analysis also suggests that the relative performance of each category of algorithms may be dependent on structural properties of the problem, such as problem flexibility and instance size. Overall, our findings challenge the prevailing assumption that policy-gradient algorithms are inherently superior for combinatorial optimization. We show instead that value-based algorithms can match or even surpass the performance of policy-gradient algorithms, suggesting that they deserve greater attention from the combinatorial optimization community. Our code is openly available at: https://github.com/AJ-Correa/Unraveling-the-Rainbow", "date": "2025-05-06", "categories": []} {"id": "2509.01886", "title": "Deep Reinforcement Learning for Drone Route Optimization in Post-Disaster Road Assessment", "abstract": "Rapid post-disaster road damage assessment is critical for effective emergency response, yet traditional optimization methods suffer from excessive computational time and require domain knowledge for algorithm design, making them unsuitable for time-sensitive disaster scenarios. This study proposes an attention-based encoder-decoder model (AEDM) for rapid drone routing decision in post-disaster road damage assessment. The method employs deep reinforcement learning to determine high-quality drone assessment routes without requiring algorithmic design knowledge. A network transformation method is developed to convert link-based routing problems into equivalent node-based formulations, while a synthetic road network generation technique addresses the scarcity of large-scale training datasets. The model is trained using policy optimization with multiple optima (POMO) with multi-task learning capabilities to handle diverse parameter combinations. Experimental results demonstrate two key strengths of AEDM: it outperforms commercial solvers by 20--71\\% and traditional heuristics by 23--35\\% in solution quality, while achieving rapid inference (1--2 seconds) versus 100--2,000 seconds for traditional methods. The model exhibits strong generalization across varying problem scales, drone numbers, and time constraints, consistently outperforming baseline methods on unseen parameter distributions and real-world road networks. The proposed method effectively balances computational efficiency with solution quality, making it particularly suitable for time-critical disaster response applications where rapid decision-making is essential for saving lives. The source code for AEDM is publicly available at https://github.com/PJ-HTU/AEDM-for-Post-disaster-road-assessment.", "date": "2025-09-02", "categories": []} {"id": "2511.21886", "title": "Bridging Planning and Execution: Multi-Agent Path Finding Under Real-World Deadlines", "abstract": "The Multi-Agent Path Finding (MAPF) problem aims to find collision-free paths for multiple agents while optimizing objectives such as the sum of costs or makespan. MAPF has wide applications in domains like automated warehouses, manufacturing systems, and airport logistics. However, most MAPF formulations assume a simplified robot model for planning, which overlooks execution-time factors such as kinodynamic constraints, communication latency, and controller variability. This gap between planning and execution is problematic for time-sensitive applications. To bridge this gap, we propose REMAP, an execution-informed MAPF planning framework that can be combined with leading search-based MAPF planners with minor changes. Our framework integrates the proposed ExecTimeNet to accurately estimate execution time based on planned paths. We demonstrate our method for solving MAPF with Real-world Deadlines (MAPF-RD) problem, where agents must reach their goals before a predefined wall-clock time. We integrate our framework with two popular MAPF methods, MAPF-LNS and CBS. Experiments show that REMAP achieves up to 20% improvement in solution quality over baseline methods (e.g., constant execution speed estimators) on benchmark maps with up to 300 agents.", "date": "2025-11-26", "categories": []} {"id": "2511.22632", "title": "Optimized Agent Shift Scheduling Using Multi-Phase Allocation Approach", "abstract": "Effective agent shift scheduling is crucial for businesses, especially in the Contact Center as a Service (CCaaS) industry, to ensure seamless operations and fulfill employee needs. Most studies utilizing mathematical model-based solutions approach the problem as a single-step process, often resulting in inefficiencies and high computational demands. In contrast, we present a multi-phase allocation method that addresses scalability and accuracy by dividing the problem into smaller sub-problems of day and shift allocation, which significantly reduces number of computational variables and allows for targeted objective functions, ultimately enhancing both efficiency and accuracy. Each subproblem is modeled as a Integer Programming Problem (IPP), with solutions sequentially feeding into the subsequent subproblem. We then apply the proposed method, using a multi-objective framework, to address the difficulties posed by peak demand scenarios such as holiday rushes, where maintaining service levels is essential despite having limited number of employees", "date": "2025-11-27", "categories": []} {"id": "2408.15041", "title": "Earth Observation Satellite Scheduling with Graph Neural Networks and Monte Carlo Tree Search", "abstract": "Earth Observation Satellite Planning (EOSP) is a difficult optimization problem with considerable practical interest. A set of requested observations must be scheduled on an agile Earth observation satellite while respecting constraints on their visibility window, as well as maneuver constraints that impose varying delays between successive observations. In addition, the problem is largely oversubscribed: there are much more candidate observations than can possibly be achieved. Therefore, one must select the set of observations that will be performed while maximizing their cumulative benefit and propose a feasible schedule for these observations. As previous work mostly focused on heuristic and iterative search algorithms, this paper presents a new technique for selecting and scheduling observations based on Graph Neural Networks (GNNs) and Deep Reinforcement Learning (DRL). GNNs are used to extract relevant information from the graphs representing instances of the EOSP, and DRL drives the search for optimal schedules. A post-learning search step based on Monte Carlo Tree Search (MCTS) is added that is able to find even better solutions. Experiments show that it is able to learn on small problem instances and generalize to larger real-world instances, with very competitive performance compared to traditional approaches.", "date": "2024-08-27", "categories": []} {"id": "2511.21107", "title": "Dynamic Stratified Contrastive Learning with Upstream Augmentation for MILP Branching", "abstract": "Mixed Integer Linear Programming (MILP) is a fundamental class of NP-hard problems that has garnered significant attention from both academia and industry. The Branch-and-Bound (B\\&B) method is the dominant approach for solving MILPs and the branching plays an important role in B\\&B methods. Neural-based learning frameworks have recently been developed to enhance branching policies and the efficiency of solving MILPs. However, these methods still struggle with semantic variation across depths, the scarcity of upstream nodes, and the costly collection of strong branching samples. To address these issues, we propose \\ours, a Dynamic \\underline{\\textbf{S}}tratified \\underline{\\textbf{C}}ontrastive Training Framework for \\underline{\\textbf{MILP}} Branching. It groups branch-and-bound nodes based on their feature distributions and trains a GCNN-based discriminative model to progressively separate nodes across groups, learning finer-grained node representations throughout the tree. To address data scarcity and imbalance at upstream nodes, we introduce an upstream-augmented MILP derivation procedure that generates both theoretically equivalent and perturbed instances. \\ours~effectively models subtle semantic differences between nodes, significantly enhancing branching accuracy and solving efficiency, particularly for upstream nodes. Extensive experiments on standard MILP benchmarks demonstrate that our method enhances branching accuracy, reduces solving time, and generalizes effectively to unseen instances.", "date": "2025-11-26", "categories": []} {"id": "2511.19544", "title": "Learning to Solve Weighted Maximum Satisfiability with a Co-Training Architecture", "abstract": "Wepropose SplitGNN, a graph neural network (GNN)-based\n approach that learns to solve weighted maximum satisfiabil ity (MaxSAT) problem. SplitGNN incorporates a co-training\n architecture consisting of supervised message passing mech anism and unsupervised solution boosting layer. A new graph\n representation called edge-splitting factor graph is proposed\n to provide more structural information for learning, which is\n based on spanning tree generation and edge classification. To\n improve the solutions on challenging and weighted instances,\n we implement a GPU-accelerated layer applying efficient\n score calculation and relaxation-based optimization. Exper iments show that SplitGNN achieves 3* faster convergence\n and better predictions compared with other GNN-based ar chitectures. More notably, SplitGNN successfully finds solu tions that outperform modern heuristic MaxSAT solvers on\n much larger and harder weighted MaxSAT benchmarks, and\n demonstrates exceptional generalization abilities on diverse\n structural instances.", "date": "2025-11-24", "categories": []} {"id": "2511.19573", "title": "Neural Tractability via Structure: Learning-Augmented Algorithms for Graph Combinatorial Optimization", "abstract": "Neural models have shown promise in solving NP-hard graph combinatorial optimization (CO) problems. Once trained, they offer fast inference and reasonably high-quality solutions for in-distribution testing instances, but they generally fall short in terms of absolute solution quality compared to classical search-based algorithms that are admittedly slower but offer optimality guarantee once search finishes.\n We propose a novel framework that combines the inference efficiency and exploratory power of neural models with the solution quality guarantee of search-based algorithms. In particular, we use parameterized algorithms (PAs) as the search component. PAs are dedicated to identifying easy instances of generally NP-hard problems, and allow for practically efficient search by exploiting structural simplicity (of the identified easy instances). Under our framework, we use parameterized analysis to identify the structurally hard parts of a CO instance. The neural model handles the hard parts by generating advisory signals based on its data-driven understanding. The PA-based search component then integrates the advisory signals to systematically and efficiently searches through the remaining structurally easy parts. Notably, our framework is agnostic to the choice of neural model and produces strictly better solutions than neural solvers alone.\n We examine our framework on multiple CO tasks. Empirical results show that it achieves superior solution quality, competitive with that of commercial solvers. Furthermore, by using the neural model only for exploratory advisory signals, our framework exhibits improved out-of-distribution generalization, addressing a key limitation of existing neural CO solvers.", "date": "2025-11-24", "categories": []} {"id": "2511.20607", "title": "Optimization of Sums of Bivariate Functions: An Introduction to Relaxation-Based Methods for the Case of Finite Domains", "abstract": "We study the optimization of functions with $n>2$ arguments that have a representation as a sum of several functions that have only $2$ of the $n$ arguments each, termed sums of bivariates, on finite domains. The complexity of optimizing sums of bivariates is shown to be NP-equivalent and it is shown that there exists free lunch in the optimization of sums of bivariates. Based on measure-valued extensions of the objective function, so-called relaxations, $\\ell^2$-approximation, and entropy-regularization, we derive several tractable problem formulations solvable with linear programming, coordinate ascent as well as with closed-form solutions. The limits of applying tractable versions of such relaxations to sums of bivariates are investigated using general results for reconstructing measures from their bivariate marginals. Experiments in which the derived algorithms are applied to random functions, vertex coloring, and signal reconstruction problems provide insights into qualitatively different function classes that can be modeled as sums of bivariates.", "date": "2025-11-25", "categories": []} {"id": "2305.15118", "title": "Fairness in Streaming Submodular Maximization over a Matroid Constraint", "abstract": "Streaming submodular maximization is a natural model for the task of selecting a representative subset from a large-scale dataset. If datapoints have sensitive attributes such as gender or race, it becomes important to enforce fairness to avoid bias and discrimination. This has spurred significant interest in developing fair machine learning algorithms. Recently, such algorithms have been developed for monotone submodular maximization under a cardinality constraint.\n In this paper, we study the natural generalization of this problem to a matroid constraint. We give streaming algorithms as well as impossibility results that provide trade-offs between efficiency, quality and fairness. We validate our findings empirically on a range of well-known real-world applications: exemplar-based clustering, movie recommendation, and maximum coverage in social networks.", "date": "2023-05-24", "categories": []} {"id": "2511.18269", "title": "A Fair OR-ML Framework for Resource Substitution in Large-Scale Networks", "abstract": "Ensuring that the right resource is available at the right location and time remains a major challenge for organizations operating large-scale logistics networks. The challenge comes from uneven demand patterns and the resulting asymmetric flow of resources across the arcs, which create persistent imbalances at the network nodes. Resource substitution among multiple, potentially composite and interchangeable, resource types is a cost-effective way to mitigate these imbalances. This leads to the resource substitution problem, which aims at determining the minimum number of resource substitutions from an initial assignment to minimize the overall network imbalance. In decentralized settings, achieving globally coordinated solutions becomes even more difficult. When substitution entails costs, effective prescriptions must also incorporate fairness and account for the individual preferences of schedulers. This paper presents a generic framework that combines operations research (OR) and machine learning (ML) to enable fair resource substitution in large networks. The OR component models and solves the resource substitution problem under a fairness lens. The ML component leverages historical data to learn schedulers' preferences, guide intelligent exploration of the decision space, and enhance computational efficiency by dynamically selecting the top-$\\kappa$ resources for each arc in the network. The framework produces a portfolio of high-quality solutions from which schedulers can select satisfactory trade-offs. The proposed framework is applied to the network of one of the largest package delivery companies in the world, which serves as the primary motivation for this research. Computational results demonstrate substantial improvements over state-of-the-art methods, including an 80% reduction in model size and a 90% decrease in execution time while preserving optimality.", "date": "2025-11-23", "categories": []} {"id": "2409.18972", "title": "Instance Configuration for Sustainable Job Shop Scheduling", "abstract": "The Job Shop Scheduling Problem (JSP) is a pivotal challenge in operations research and is essential for evaluating the effectiveness and performance of scheduling algorithms. Scheduling problems are a crucial domain in combinatorial optimization, where resources (machines) are allocated to job tasks to minimize the completion time (makespan) alongside other objectives like energy consumption. This research delves into the intricacies of JSP, focusing on optimizing performance metrics and minimizing energy consumption while considering various constraints such as deadlines and release dates. Recognizing the multi-dimensional nature of benchmarking in JSP, this study underscores the significance of reference libraries and datasets like JSPLIB in enriching algorithm evaluation. The research highlights the importance of problem instance characteristics, including job and machine numbers, processing times, and machine availability, emphasizing the complexities introduced by energy consumption considerations.\n An innovative instance configurator is proposed, equipped with parameters such as the number of jobs, machines, tasks, and speeds, alongside distributions for processing times and energy consumption. The generated instances encompass various configurations, reflecting real-world scenarios and operational constraints. These instances facilitate comprehensive benchmarking and evaluation of scheduling algorithms, particularly in contexts of energy efficiency. A comprehensive set of 500 test instances has been generated and made publicly available, promoting further research and benchmarking in JSP. These instances enable robust analyses and foster collaboration in developing advanced, energy-efficient scheduling solutions by providing diverse scenarios.", "date": "2024-09-12", "categories": []} {"id": "2506.02392", "title": "Improving Generalization of Neural Combinatorial Optimization for Vehicle Routing Problems via Test-Time Projection Learning", "abstract": "Neural Combinatorial Optimization (NCO) has emerged as a promising learning-based paradigm for addressing Vehicle Routing Problems (VRPs) by minimizing the need for extensive manual engineering. While existing NCO methods, trained on small-scale instances (e.g., 100 nodes), have demonstrated considerable success on problems of similar scale, their performance significantly degrades when applied to large-scale scenarios. This degradation arises from the distributional shift between training and testing data, rendering policies learned on small instances ineffective for larger problems. To overcome this limitation, we introduce a novel learning framework driven by Large Language Models (LLMs). This framework learns a projection between the training and testing distributions, which is then deployed to enhance the scalability of the NCO model. Notably, unlike prevailing techniques that necessitate joint training with the neural network, our approach operates exclusively during the inference phase, obviating the need for model retraining. Extensive experiments demonstrate that our method enables a backbone model (trained on 100-node instances) to achieve superior performance on large-scale Traveling Salesman Problem (TSP) and Capacitated Vehicle Routing Problem (CVRP) of up to 100K nodes from diverse distributions.", "date": "2025-06-03", "categories": []} {"id": "2511.16637", "title": "Faster Certified Symmetry Breaking Using Orders With Auxiliary Variables", "abstract": "Symmetry breaking is a crucial technique in modern combinatorial solving, but it is difficult to be sure it is implemented correctly. The most successful approach to deal with bugs is to make solvers certifying, so that they output not just a solution, but also a mathematical proof of correctness in a standard format, which can then be checked by a formally verified checker. This requires justifying symmetry reasoning within the proof, but developing efficient methods for this has remained a long-standing open challenge. A fully general approach was recently proposed by Bogaerts et al. (2023), but it relies on encoding lexicographic orders with big integers, which quickly becomes infeasible for large symmetries. In this work, we develop a method for instead encoding orders with auxiliary variables. We show that this leads to orders-of-magnitude speed-ups in both theory and practice by running experiments on proof logging and checking for SAT symmetry breaking using the state-of-the-art satsuma symmetry breaker and the VeriPB proof checking toolchain.", "date": "2025-11-20", "categories": []} {"id": "2502.06440", "title": "SIGMA: Sheaf-Informed Geometric Multi-Agent Pathfinding", "abstract": "The Multi-Agent Path Finding (MAPF) problem aims to determine the shortest and collision-free paths for multiple agents in a known, potentially obstacle-ridden environment. It is the core challenge for robotic deployments in large-scale logistics and transportation. Decentralized learning-based approaches have shown great potential for addressing the MAPF problems, offering more reactive and scalable solutions. However, existing learning-based MAPF methods usually rely on agents making decisions based on a limited field of view (FOV), resulting in short-sighted policies and inefficient cooperation in complex scenarios. There, a critical challenge is to achieve consensus on potential movements between agents based on limited observations and communications. To tackle this challenge, we introduce a new framework that applies sheaf theory to decentralized deep reinforcement learning, enabling agents to learn geometric cross-dependencies between each other through local consensus and utilize them for tightly cooperative decision-making. In particular, sheaf theory provides a mathematical proof of conditions for achieving global consensus through local observation. Inspired by this, we incorporate a neural network to approximately model the consensus in latent space based on sheaf theory and train it through self-supervised learning. During the task, in addition to normal features for MAPF as in previous works, each agent distributedly reasons about a learned consensus feature, leading to efficient cooperation on pathfinding and collision avoidance. As a result, our proposed method demonstrates significant improvements over state-of-the-art learning-based MAPF planners, especially in relatively large and complex scenarios, demonstrating its superiority over baselines in various simulations and real-world robot experiments. The code is available at https://github.com/marmotlab/SIGMA", "date": "2025-02-10", "categories": []} {"id": "2508.02503", "title": "OptiHive: Ensemble Selection for LLM-Based Optimization via Statistical Modeling", "abstract": "LLM-based solvers have emerged as a promising means of automating problem modeling and solving. However, they remain unreliable and often depend on iterative repair loops that result in significant latency. We introduce OptiHive, a framework that enhances any solver-generation pipeline to produce higher-quality solvers from natural-language descriptions of optimization problems. OptiHive uses a single batched generation to produce diverse components (solvers, problem instances, and validation tests) and filters out erroneous components to ensure fully interpretable outputs. Accounting for the imperfection of the generated components, we employ a statistical model to infer their true performance, enabling principled uncertainty quantification and solver selection. On tasks ranging from traditional optimization problems to challenging variants of the Multi-Depot Vehicle Routing Problem, OptiHive significantly outperforms baselines, increasing the optimality rate from 5% to 92% on the most complex problems.", "date": "2025-08-04", "categories": []} {"id": "2508.07015", "title": "Efficient and Reliable Hitting-Set Computations for the Implicit Hitting Set Approach", "abstract": "The implicit hitting set (IHS) approach offers a general framework for solving computationally hard combinatorial optimization problems declaratively. IHS iterates between a decision oracle used for extracting sources of inconsistency and an optimizer for computing so-called hitting sets (HSs) over the accumulated sources of inconsistency. While the decision oracle is language-specific, the optimizers is usually instantiated through integer programming.\n We explore alternative algorithmic techniques for hitting set optimization based on different ways of employing pseudo-Boolean (PB) reasoning as well as stochastic local search. We extensively evaluate the practical feasibility of the alternatives in particular in the context of pseudo-Boolean (0-1 IP) optimization as one of the most recent instantiations of IHS. Highlighting a trade-off between efficiency and reliability, while a commercial IP solver turns out to remain the most effective way to instantiate HS computations, it can cause correctness issues due to numerical instability; in fact, we show that exact HS computations instantiated via PB reasoning can be made competitive with a numerically exact IP solver. Furthermore, the use of PB reasoning as a basis for HS computations allows for obtaining certificates for the correctness of IHS computations, generally applicable to any IHS instantiation in which reasoning in the declarative language at hand can be captured in the PB-based proof format we employ.", "date": "2025-08-09", "categories": []} {"id": "2511.12986", "title": "Learning Branching Policies for MILPs with Proximal Policy Optimization", "abstract": "Branch-and-Bound (B\\&B) is the dominant exact solution method for Mixed Integer Linear Programs (MILP), yet its exponential time complexity poses significant challenges for large-scale instances. The growing capabilities of machine learning have spurred efforts to improve B\\&B by learning data-driven branching policies. However, most existing approaches rely on Imitation Learning (IL), which tends to overfit to expert demonstrations and struggles to generalize to structurally diverse or unseen instances. In this work, we propose Tree-Gate Proximal Policy Optimization (TGPPO), a novel framework that employs Proximal Policy Optimization (PPO), a Reinforcement Learning (RL) algorithm, to train a branching policy aimed at improving generalization across heterogeneous MILP instances. Our approach builds on a parameterized state space representation that dynamically captures the evolving context of the search tree. Empirical evaluations show that TGPPO often outperforms existing learning-based policies in terms of reducing the number of nodes explored and improving p-Primal-Dual Integrals (PDI), particularly in out-of-distribution instances. These results highlight the potential of RL to develop robust and adaptable branching strategies for MILP solvers.", "date": "2025-11-17", "categories": []} {"id": "2511.13214", "title": "Learning to Solve Resource-Constrained Project Scheduling Problems with Duration Uncertainty using Graph Neural Networks", "abstract": "The Resource-Constrained Project Scheduling Problem (RCPSP) is a classical scheduling problem that has received significant attention due to of its numerous applications in industry. However, in practice, task durations are subject to uncertainty that must be considered in order to propose resilient scheduling. In this paper, we address the RCPSP variant with uncertain tasks duration (modeled using known probabilities) and aim to minimize the overall expected project duration. Our objective is to produce a baseline schedule that can be reused multiple times in an industrial setting regardless of the actual duration scenario. We leverage Graph Neural Networks in conjunction with Deep Reinforcement Learning (DRL) to develop an effective policy for task scheduling. This policy operates similarly to a priority dispatch rule and is paired with a Serial Schedule Generation Scheme to produce a schedule. Our empirical evaluation on standard benchmarks demonstrates the approach's superiority in terms of performance and its ability to generalize. The developed framework, Wheatley, is made publicly available online to facilitate further research and reproducibility.", "date": "2025-11-17", "categories": []} {"id": "2511.10804", "title": "Discounted Cuts: A Stackelberg Approach to Network Disruption", "abstract": "We study a Stackelberg variant of the classical Most Vital Links problem, modeled as a one-round adversarial game between an attacker and a defender. The attacker strategically removes up to $k$ edges from a flow network to maximally disrupt flow between a source $s$ and a sink $t$, after which the defender optimally reroutes the remaining flow. To capture this attacker--defender interaction, we introduce a new mathematical model of discounted cuts, in which the cost of a cut is evaluated by excluding its $k$ most expensive edges. This model generalizes the Most Vital Links problem and uncovers novel algorithmic and complexity-theoretic properties.\n We develop a unified algorithmic framework for analyzing various forms of discounted cut problems, including minimizing or maximizing the cost of a cut under discount mechanisms that exclude either the $k$ most expensive or the $k$ cheapest edges. While most variants are NP-complete on general graphs, our main result establishes polynomial-time solvability for all discounted cut problems in our framework when the input is restricted to bounded-genus graphs, a relevant class that includes many real-world networks such as transportation and infrastructure networks. With this work, we aim to open collaborative bridges between artificial intelligence, algorithmic game theory, and operations research.", "date": "2025-11-13", "categories": []} {"id": "2511.09563", "title": "An Efficient and Almost Optimal Solver for the Joint Routing-Assignment Problem via Partial JRA and Large-{\\alpha} Optimization", "abstract": "The Joint Routing-Assignment (JRA) optimization problem simultaneously determines the assignment of items to placeholders and a Hamiltonian cycle that visits each node pair exactly once, with the objective of minimizing total travel cost. Previous studies introduced an exact mixed-integer programming (MIP) solver, along with datasets and a Gurobi implementation, showing that while the exact approach guarantees optimality, it becomes computationally inefficient for large-scale instances. To overcome this limitation, heuristic methods based on merging algorithms and shaking procedures were proposed, achieving solutions within approximately 1% deviation from the optimum. This work presents a novel and more efficient approach that attains high-accuracy, near-optimal solutions for large-scale JRA problems. The proposed method introduces a Partial Path Reconstructon (PPR) solver that first identifies key item-placeholder pairs to form a reduced subproblem, which is solved efficiently to refine the global solution. Using this PJAR framework, the initial heuristic merging solutions can be further improved, reducing the deviation by half. Moreover, the solution can be iteratively polished with PPR based solver along the optimization path to yield highly accurate tours. Additionally, a global Large-{\\alpha} constraint is incorporated into the JRA model to further enhance solution optimality. Experimental evaluations on benchmark datasets with n = 300, 500, and 1000 demonstrate that the proposed method consistently delivers almost optimal solutions, achieving an average deviation of 0.00% from the ground truth while maintaining high computational efficiency. Beyond the JRA problem, the proposed framework and methodologies exhibit strong potential for broader applications. The Framework can be applied to TSP and related optimization problems.", "date": "2025-11-07", "categories": []} {"id": "2511.09570", "title": "Variable Neighborhood Search for the Electric Vehicle Routing Problem", "abstract": "The Electric Vehicle Routing Problem (EVRP) extends the classical Vehicle Routing Problem (VRP) to reflect the growing use of electric and hybrid vehicles in logistics. Due to the variety of constraints considered in the literature, comparing approaches across different problem variants remains challenging. A minimalistic variant of the EVRP, known as the Capacitated Green Vehicle Routing Problem (CGVRP), was the focus of the CEC-12 competition held during the 2020 IEEE World Congress on Computational Intelligence. This paper presents the competition-winning approach, based on the Variable Neighborhood Search (VNS) metaheuristic. The method achieves the best results on the full competition dataset and also outperforms a more recent algorithm published afterward.", "date": "2025-11-11", "categories": []} {"id": "2511.10272", "title": "Bidirectional Bounded-Suboptimal Heuristic Search with Consistent Heuristics", "abstract": "Recent advancements in bidirectional heuristic search have yielded significant theoretical insights and novel algorithms. While most previous work has concentrated on optimal search methods, this paper focuses on bounded-suboptimal bidirectional search, where a bound on the suboptimality of the solution cost is specified. We build upon the state-of-the-art optimal bidirectional search algorithm, BAE*, designed for consistent heuristics, and introduce several variants of BAE* specifically tailored for the bounded-suboptimal context. Through experimental evaluation, we compare the performance of these new variants against other bounded-suboptimal bidirectional algorithms as well as the standard weighted A* algorithm. Our results demonstrate that each algorithm excels under distinct conditions, highlighting the strengths and weaknesses of each approach.", "date": "2025-11-13", "categories": []} {"id": "2511.09093", "title": "Factorization-in-Loop: Proximal Fill-in Minimization for Sparse Matrix Reordering", "abstract": "Fill-ins are new nonzero elements in the summation of the upper and lower triangular factors generated during LU factorization. For large sparse matrices, they will increase the memory usage and computational time, and be reduced through proper row or column arrangement, namely matrix reordering. Finding a row or column permutation with the minimal fill-ins is NP-hard, and surrogate objectives are designed to derive fill-in reduction permutations or learn a reordering function. However, there is no theoretical guarantee between the golden criterion and these surrogate objectives. Here we propose to learn a reordering network by minimizing \\(l_1\\) norm of triangular factors of the reordered matrix to approximate the exact number of fill-ins. The reordering network utilizes a graph encoder to predict row or column node scores. For inference, it is easy and fast to derive the permutation from sorting algorithms for matrices. For gradient based optimization, there is a large gap between the predicted node scores and resultant triangular factors in the optimization objective. To bridge the gap, we first design two reparameterization techniques to obtain the permutation matrix from node scores. The matrix is reordered by multiplying the permutation matrix. Then we introduce the factorization process into the objective function to arrive at target triangular factors. The overall objective function is optimized with the alternating direction method of multipliers and proximal gradient descent. Experimental results on benchmark sparse matrix collection SuiteSparse show the fill-in number and LU factorization time reduction of our proposed method is 20% and 17.8% compared with state-of-the-art baselines.", "date": "2025-11-12", "categories": []} {"id": "2511.09209", "title": "CoCo-MILP: Inter-Variable Contrastive and Intra-Constraint Competitive MILP Solution Prediction", "abstract": "Mixed-Integer Linear Programming (MILP) is a cornerstone of combinatorial optimization, yet solving large-scale instances remains a significant computational challenge. Recently, Graph Neural Networks (GNNs) have shown promise in accelerating MILP solvers by predicting high-quality solutions. However, we identify that existing methods misalign with the intrinsic structure of MILP problems at two levels. At the leaning objective level, the Binary Cross-Entropy (BCE) loss treats variables independently, neglecting their relative priority and yielding plausible logits. At the model architecture level, standard GNN message passing inherently smooths the representations across variables, missing the natural competitive relationships within constraints. To address these challenges, we propose CoCo-MILP, which explicitly models inter-variable Contrast and intra-constraint Competition for advanced MILP solution prediction. At the objective level, CoCo-MILP introduces the Inter-Variable Contrastive Loss (VCL), which explicitly maximizes the embedding margin between variables assigned one versus zero. At the architectural level, we design an Intra-Constraint Competitive GNN layer that, instead of homogenizing features, learns to differentiate representations of competing variables within a constraint, capturing their exclusionary nature. Experimental results on standard benchmarks demonstrate that CoCo-MILP significantly outperforms existing learning-based approaches, reducing the solution gap by up to 68.12% compared to traditional solvers. Our code is available at https://github.com/happypu326/CoCo-MILP.", "date": "2025-11-12", "categories": []} {"id": "2511.09261", "title": "A Distributed Training Architecture For Combinatorial Optimization", "abstract": "In recent years, graph neural networks (GNNs) have been widely applied in tackling combinatorial optimization problems. However, existing methods still suffer from limited accuracy when addressing that on complex graphs and exhibit poor scalability, since full training requires loading the whole adjacent matrix and all embeddings at a time, the it may results in out of memory of a single machine. This limitation significantly restricts their applicability to large-scale scenarios. To address these challenges, we propose a distributed GNN-based training framework for combinatorial optimization. In details, firstly, large graph is partition into several small subgraphs. Then the individual subgraphs are full trained, providing a foundation for efficient local optimization. Finally, reinforcement learning (RL) are employed to take actions according to GNN output, to make sure the restrictions between cross nodes can be learned. Extensive experiments are conducted on both real large-scale social network datasets (e.g., Facebook, Youtube) and synthetically generated high-complexity graphs, which demonstrate that our framework outperforms state-of-the-art approaches in both solution quality and computational efficiency. Moreover, the experiments on large graph instances also validate the scalability of the model.", "date": "2025-11-12", "categories": []} {"id": "2511.07724", "title": "A Ranking-Based Optimization Algorithm for the Vehicle Relocation Problem in Car Sharing Services", "abstract": "The paper addresses the Vehicle Relocation Problem in free-floating car-sharing services by presenting a solution focused on strategies for repositioning vehicles and transferring personnel with the use of scooters. Our method begins by dividing the service area into zones that group regions with similar temporal patterns of vehicle presence and service demand, allowing the application of discrete optimization methods. In the next stage, we propose a fast ranking-based algorithm that makes its decisions on the basis of the number of cars available in each zone, the projected probability density of demand, and estimated trip durations. The experiments were carried out on the basis of real-world data originating from a major car-sharing service operator in Poland. The results of this algorithm are evaluated against scenarios without optimization that constitute a baseline and compared with the results of an exact algorithm to solve the Mixed Integer Programming (MIP) model. As performance metrics, the total travel time was used. Under identical conditions (number of vehicles, staff, and demand distribution), the average improvements with respect to the baseline of our algorithm and MIP solver were equal to 8.44\\% and 19.6\\% correspondingly. However, it should be noted that the MIP model also mimicked decisions on trip selection, which are excluded by current services business rules. The analysis of results suggests that, depending on the size of the workforce, the application of the proposed solution allows for improving performance metrics by roughly 3%-10%.", "date": "2025-11-11", "categories": []} {"id": "2511.07737", "title": "TurboSAT: Gradient-Guided Boolean Satisfiability Accelerated on GPU-CPU Hybrid System", "abstract": "While accelerated computing has transformed many domains of computing, its impact on logical reasoning, specifically Boolean satisfiability (SAT), remains limited. State-of-the-art SAT solvers rely heavily on inherently sequential conflict-driven search algorithms that offer powerful heuristics but limit the amount of parallelism that could otherwise enable significantly more scalable SAT solving. Inspired by neural network training, we formulate the SAT problem as a binarized matrix-matrix multiplication layer that could be optimized using a differentiable objective function. Enabled by this encoding, we combine the strengths of parallel differentiable optimization and sequential search to accelerate SAT on a hybrid GPU-CPU system. In this system, the GPUs leverage parallel differentiable solving to rapidly evaluate SAT clauses and use gradients to stochastically explore the solution space and optimize variable assignments. Promising partial assignments generated by the GPUs are post-processed on many CPU threads which exploit conflict-driven sequential search to further traverse the solution subspaces and identify complete assignments. Prototyping the hybrid solver on an NVIDIA DGX GB200 node, our solver achieves runtime speedups up to over 200x when compared to a state-of-the-art CPU-based solver on public satisfiable benchmark problems from the SAT Competition.", "date": "2025-11-11", "categories": []} {"id": "2511.08275", "title": "Bi-Objective Evolutionary Optimization for Large-Scale Open Pit Mine Scheduling Problem under Uncertainty with Chance Constraints", "abstract": "The open-pit mine scheduling problem (OPMSP) is a complex, computationally expensive process in long-term mine planning, constrained by operational and geological dependencies. Traditional deterministic approaches often ignore geological uncertainty, leading to suboptimal and potentially infeasible production schedules. Chance constraints allow modeling of stochastic components by ensuring probabilistic constraints are satisfied with high probability. This paper presents a bi-objective formulation of the OPMSP that simultaneously maximizes expected net present value and minimizes scheduling risk, independent of the confidence level required for the constraint. Solutions are represented using integer encoding, inherently satisfying reserve constraints. We introduce a domain-specific greedy randomized initialization and a precedence-aware period-swap mutation operator. We integrate these operators into three multi-objective evolutionary algorithms: the global simple evolutionary multi-objective optimizer (GSEMO), a mutation-only variant of multi-objective evolutionary algorithm based on decomposition (MOEA/D), and non-dominated sorting genetic algorithm II (NSGA-II). We compare our bi-objective formulation against the single-objective approach, which depends on a specific confidence level, by analyzing mine deposits consisting of up to 112 687 blocks. Results demonstrate that the proposed bi-objective formulation yields more robust and balanced trade-offs between economic value and risk compared to single-objective, confidence-dependent approach.", "date": "2025-11-11", "categories": []} {"id": "2511.05872", "title": "Adaptation and Fine-tuning with TabPFN for Travelling Salesman Problem", "abstract": "Tabular Prior-Data Fitted Network (TabPFN) is a foundation model designed for small to medium-sized tabular data, which has attracted much attention recently. This paper investigates the application of TabPFN in Combinatorial Optimization (CO) problems. The aim is to lessen challenges in time and data-intensive training requirements often observed in using traditional methods including exact and heuristic algorithms, Machine Learning (ML)-based models, to solve CO problems. Proposing possibly the first ever application of TabPFN for such a purpose, we adapt and fine-tune the TabPFN model to solve the Travelling Salesman Problem (TSP), one of the most well-known CO problems. Specifically, we adopt the node-based approach and the node-predicting adaptation strategy to construct the entire TSP route. Our evaluation with varying instance sizes confirms that TabPFN requires minimal training, adapts to TSP using a single sample, performs better generalization across varying TSP instance sizes, and reduces performance degradation. Furthermore, the training process with adaptation and fine-tuning is completed within minutes. The methodology leads to strong solution quality even without post-processing and achieves performance comparable to other models with post-processing refinement. Our findings suggest that the TabPFN model is a promising approach to solve structured and CO problems efficiently under training resource constraints and rapid deployment requirements.", "date": "2025-11-08", "categories": []} {"id": "2511.05265", "title": "An End-to-End Deep Reinforcement Learning Approach for Solving the Traveling Salesman Problem with Drones", "abstract": "The emergence of truck-drone collaborative systems in last-mile logistics has positioned the Traveling Salesman Problem with Drones (TSP-D) as a pivotal extension of classical routing optimization, where synchronized vehicle coordination promises substantial operational efficiency and reduced environmental impact, yet introduces NP-hard combinatorial complexity beyond the reach of conventional optimization paradigms. Deep reinforcement learning offers a theoretically grounded framework to address TSP-D's inherent challenges through self-supervised policy learning and adaptive decision-making. This study proposes a hierarchical Actor-Critic deep reinforcement learning framework for solving the TSP-D problem. The architecture consists of two primary components: a Transformer-inspired encoder and an efficient Minimal Gated Unit decoder. The encoder incorporates a novel, optimized k-nearest neighbors sparse attention mechanism specifically for focusing on relevant spatial relationships, further enhanced by the integration of global node features. The Minimal Gated Unit decoder processes these encoded representations to efficiently generate solution sequences. The entire framework operates within an asynchronous advantage actor-critic paradigm. Experimental results show that, on benchmark TSP-D instances of various scales (N=10 to 100), the proposed model can obtain competitive or even superior solutions in shorter average computation times compared to high-performance heuristic algorithms and existing reinforcement learning methods. Moreover, compared to advanced reinforcement learning algorithm benchmarks, the proposed framework significantly reduces the total training time required while achieving superior final performance, highlighting its notable advantage in training efficiency.", "date": "2025-11-07", "categories": []} {"id": "2511.04183", "title": "A Reinforced Evolution-Based Approach to Multi-Resource Load Balancing", "abstract": "This paper presents a reinforced genetic approach to a defined d-resource system optimization problem. The classical evolution schema was ineffective due to a very strict feasibility function in the studied problem. Hence, the presented strategy has introduced several modifications and adaptations to standard genetic routines, e.g.: a migration operator which is an analogy to the biological random genetic drift.", "date": "2025-11-06", "categories": []} {"id": "2506.02935", "title": "MTL-KD: Multi-Task Learning Via Knowledge Distillation for Generalizable Neural Vehicle Routing Solver", "abstract": "Multi-Task Learning (MTL) in Neural Combinatorial Optimization (NCO) is a promising approach to train a unified model capable of solving multiple Vehicle Routing Problem (VRP) variants. However, existing Reinforcement Learning (RL)-based multi-task methods can only train light decoder models on small-scale problems, exhibiting limited generalization ability when solving large-scale problems. To overcome this limitation, this work introduces a novel multi-task learning method driven by knowledge distillation (MTL-KD), which enables the efficient training of heavy decoder models with strong generalization ability. The proposed MTL-KD method transfers policy knowledge from multiple distinct RL-based single-task models to a single heavy decoder model, facilitating label-free training and effectively improving the model's generalization ability across diverse tasks. In addition, we introduce a flexible inference strategy termed Random Reordering Re-Construction (R3C), which is specifically adapted for diverse VRP tasks and further boosts the performance of the multi-task model. Experimental results on 6 seen and 10 unseen VRP variants with up to 1000 nodes indicate that our proposed method consistently achieves superior performance on both uniform and real-world benchmarks, demonstrating robust generalization abilities.", "date": "2025-06-03", "categories": []} {"id": "2506.07952", "title": "Discrete and Continuous Difference of Submodular Minimization", "abstract": "Submodular functions, defined on continuous or discrete domains, arise in numerous applications. We study the minimization of the difference of two submodular (DS) functions, over both domains, extending prior work restricted to set functions. We show that all functions on discrete domains and all smooth functions on continuous domains are DS. For discrete domains, we observe that DS minimization is equivalent to minimizing the difference of two convex (DC) functions, as in the set function case. We propose a novel variant of the DC Algorithm (DCA) and apply it to the resulting DC Program, obtaining comparable theoretical guarantees as in the set function case. The algorithm can be applied to continuous domains via discretization. Experiments demonstrate that our method outperforms baselines in integer compressive sensing and integer least squares.", "date": "2025-06-09", "categories": []} {"id": "2511.02048", "title": "Finding Probably Approximate Optimal Solutions by Training to Estimate the Optimal Values of Subproblems", "abstract": "The paper is about developing a solver for maximizing a real-valued function of binary variables. The solver relies on an algorithm that estimates the optimal objective-function value of instances from the underlying distribution of objectives and their respective sub-instances. The training of the estimator is based on an inequality that facilitates the use of the expected total deviation from optimality conditions as a loss function rather than the objective-function itself. Thus, it does not calculate values of policies, nor does it rely on solved instances.", "date": "2025-11-03", "categories": []} {"id": "2511.02254", "title": "Fast Approximation Algorithm for Non-Monotone DR-submodular Maximization under Size Constraint", "abstract": "This work studies the non-monotone DR-submodular Maximization over a ground set of $n$ subject to a size constraint $k$. We propose two approximation algorithms for solving this problem named FastDrSub and FastDrSub++. FastDrSub offers an approximation ratio of $0.044$ with query complexity of $O(n \\log(k))$. The second one, FastDrSub++, improves upon it with a ratio of $1/4-\\epsilon$ within query complexity of $(n \\log k)$ for an input parameter $\\epsilon >0$. Therefore, our proposed algorithms are the first constant-ratio approximation algorithms for the problem with the low complexity of $O(n \\log(k))$.\n Additionally, both algorithms are experimentally evaluated and compared against existing state-of-the-art methods, demonstrating their effectiveness in solving the Revenue Maximization problem with DR-submodular objective function. The experimental results show that our proposed algorithms significantly outperform existing approaches in terms of both query complexity and solution quality.", "date": "2025-11-04", "categories": []} {"id": "2511.02331", "title": "RoME: Domain-Robust Mixture-of-Experts for MILP Solution Prediction across Domains", "abstract": "Mixed-Integer Linear Programming (MILP) is a fundamental and powerful framework for modeling complex optimization problems across diverse domains. Recently, learning-based methods have shown great promise in accelerating MILP solvers by predicting high-quality solutions. However, most existing approaches are developed and evaluated in single-domain settings, limiting their ability to generalize to unseen problem distributions. This limitation poses a major obstacle to building scalable and general-purpose learning-based solvers. To address this challenge, we introduce RoME, a domain-Robust Mixture-of-Experts framework for predicting MILP solutions across domains. RoME dynamically routes problem instances to specialized experts based on learned task embeddings. The model is trained using a two-level distributionally robust optimization strategy: inter-domain to mitigate global shifts across domains, and intra-domain to enhance local robustness by introducing perturbations on task embeddings. We reveal that cross-domain training not only enhances the model's generalization capability to unseen domains but also improves performance within each individual domain by encouraging the model to capture more general intrinsic combinatorial patterns. Specifically, a single RoME model trained on three domains achieves an average improvement of 67.7% then evaluated on five diverse domains. We further test the pretrained model on MIPLIB in a zero-shot setting, demonstrating its ability to deliver measurable performance gains on challenging real-world instances where existing learning-based approaches often struggle to generalize.", "date": "2025-11-04", "categories": []} {"id": "2511.02406", "title": "Arithmetic Circuits and Neural Networks for Regular Matroids", "abstract": "We prove that there exist uniform $(+,\\times,/)$-circuits of size $O(n^3)$ to compute the basis generating polynomial of regular matroids on $n$ elements. By tropicalization, this implies that there exist uniform $(\\max,+,-)$-circuits and ReLU neural networks of the same size for weighted basis maximization of regular matroids. As a consequence in linear programming theory, we obtain a first example where taking the difference of two extended formulations can be more efficient than the best known individual extended formulation of size $O(n^6)$ by Aprile and Fiorini. Such differences have recently been introduced as virtual extended formulations. The proof of our main result relies on a fine-tuned version of Seymour's decomposition of regular matroids which allows us to identify and maintain graphic substructures to which we can apply a local version of the star-mesh transformation.", "date": "2025-11-04", "categories": []} {"id": "2303.13773", "title": "Graph Neural Networks for the Offline Nanosatellite Task Scheduling Problem", "abstract": "This study investigates how to schedule nanosatellite tasks more efficiently using Graph Neural Networks (GNNs). In the Offline Nanosatellite Task Scheduling (ONTS) problem, the goal is to find the optimal schedule for tasks to be carried out in orbit while taking into account Quality-of-Service (QoS) considerations such as priority, minimum and maximum activation events, execution time-frames, periods, and execution windows, as well as constraints on the satellite's power resources and the complexity of energy harvesting and management. The ONTS problem has been approached using conventional mathematical formulations and exact methods, but their applicability to challenging cases of the problem is limited. This study examines the use of GNNs in this context, which has been effectively applied to optimization problems such as the traveling salesman, scheduling, and facility placement problems. More specifically, we investigate whether GNNs can learn the complex structure of the ONTS problem with respect to feasibility and optimality of candidate solutions. Furthermore, we evaluate using GNN-based heuristic solutions to provide better solutions (w.r.t. the objective value) to the ONTS problem and reduce the optimization cost. Our experiments show that GNNs are not only able to learn feasibility and optimality for instances of the ONTS problem, but they can generalize to harder instances than those seen during training. Furthermore, the GNN-based heuristics improved the expected objective value of the best solution found under the time limit in 45%, and reduced the expected time to find a feasible solution in 35%, when compared to the SCIP (Solving Constraint Integer Programs) solver in its off-the-shelf configuration", "date": "2023-03-24", "categories": []} {"id": "2406.16424", "title": "Memory-Enhanced Neural Solvers for Routing Problems", "abstract": "Routing Problems are central to many real-world applications, yet remain challenging due to their (NP-)hard nature. Amongst existing approaches, heuristics often offer the best trade-off between quality and scalability, making them suitable for industrial use. While Reinforcement Learning (RL) offers a flexible framework for designing heuristics, its adoption over handcrafted heuristics remains incomplete. Existing learned methods still lack the ability to adapt to specific instances and fully leverage the available computational budget. Current best methods either rely on a collection of pre-trained policies, or on RL fine-tuning; hence failing to fully utilize newly available information within the constraints of the budget. In response, we present MEMENTO, an approach that leverages memory to improve the search of neural solvers at inference. MEMENTO leverages online data collected across repeated attempts to dynamically adjust the action distribution based on the outcome of previous decisions. We validate its effectiveness on the Traveling Salesman and Capacitated Vehicle Routing problems, demonstrating its superiority over tree-search and policy-gradient fine-tuning; and showing that it can be zero-shot combined with diversity-based solvers. We successfully train all RL auto-regressive solvers on large instances, and verify MEMENTO's scalability and data-efficiency: pushing the state-of-the-art on 11 out of 12 evaluated tasks.", "date": "2024-06-24", "categories": []} {"id": "2507.19109", "title": "Pareto-NRPA: A Novel Monte-Carlo Search Algorithm for Multi-Objective Optimization", "abstract": "We introduce Pareto-NRPA, a new Monte-Carlo algorithm designed for multi-objective optimization problems over discrete search spaces. Extending the Nested Rollout Policy Adaptation (NRPA) algorithm originally formulated for single-objective problems, Pareto-NRPA generalizes the nested search and policy update mechanism to multi-objective optimization. The algorithm uses a set of policies to concurrently explore different regions of the solution space and maintains non-dominated fronts at each level of search. Policy adaptation is performed with respect to the diversity and isolation of sequences within the Pareto front. We benchmark Pareto-NRPA on two classes of problems: a novel bi-objective variant of the Traveling Salesman Problem with Time Windows problem (MO-TSPTW), and a neural architecture search task on well-known benchmarks. Results demonstrate that Pareto-NRPA achieves competitive performance against state-of-the-art multi-objective algorithms, both in terms of convergence and diversity of solutions. Particularly, Pareto-NRPA strongly outperforms state-of-the-art evolutionary multi-objective algorithms on constrained search spaces. To our knowledge, this work constitutes the first adaptation of NRPA to the multi-objective setting.", "date": "2025-07-25", "categories": []} {"id": "2508.06906", "title": "Machine Learning Algorithms for Improving Exact Classical Solvers in Mixed Integer Continuous Optimization", "abstract": "Integer and mixed-integer nonlinear programming (INLP, MINLP) are central to logistics, energy, and scheduling, but remain computationally challenging. This survey examines how machine learning and reinforcement learning can enhance exact optimization methods-particularly branch-and-bound (BB)-without compromising global optimality. We cover discrete, continuous, and mixed-integer formulations, and highlight applications such as vehicle routing, hydropower planning, and crew scheduling. We introduce a unified BB framework that embeds learning-based strategies into branching, cut selection, node ordering, and parameter control. Classical algorithms are augmented using supervised, imitation, and reinforcement learning models to accelerate convergence while maintaining correctness. We conclude with a taxonomy of learning methods by solver class and learning paradigm, and outline open challenges in generalization, hybridization, and scaling intelligent solvers.", "date": "2025-08-09", "categories": []} {"id": "2511.00015", "title": "Sorting by Strip Swaps is NP-Hard", "abstract": "We show that \\emph{Sorting by Strip Swaps} (SbSS) is NP-hard by a polynomial reduction of \\emph{Block Sorting}. The key idea is a local gadget, a \\emph{cage}, that replaces every decreasing adjacency $(a_i,a_{i+1})$ by a guarded triple $a_i,m_i,a_{i+1}$ enclosed by guards $L_i,U_i$, so the only decreasing adjacencies are the two inside the cage. Small \\emph{hinge} gadgets couple adjacent cages that share an element and enforce that a strip swap that removes exactly two adjacencies corresponds bijectively to a block move that removes exactly one decreasing adjacency in the source permutation. This yields a clean equivalence between exact SbSS schedules and perfect block schedules, establishing NP-hardness.", "date": "2025-10-20", "categories": []} {"id": "2511.00547", "title": "Efficient Generation of Binary Magic Squares", "abstract": "We propose a simple algorithm for generating Binary Magic Squares (BMS), i.e., square binary matrices where the sum of all rows and all columns are equal. We show by induction that our algorithm always returns valid BMS with optimal theoretical complexity. We then extend our study to non-square Binary Magic Squares, formalize conditions on the sum of rows and columns for these BMS to exist, and show that a slight variant of our first algorithm can generate provably generate them. Finally, we publicly release two implementations of our algorithm as Python packages, including one that can generate several BMS in parallel using GPU acceleration.", "date": "2025-11-01", "categories": []} {"id": "2511.00869", "title": "Fast Stochastic Greedy Algorithm for $k$-Submodular Cover Problem", "abstract": "We study the $k$-Submodular Cover ($kSC$) problem, a natural generalization of the classical Submodular Cover problem that arises in artificial intelligence and combinatorial optimization tasks such as influence maximization, resource allocation, and sensor placement. Existing algorithms for $\\kSC$ often provide weak approximation guarantees or incur prohibitively high query complexity. To overcome these limitations, we propose a \\textit{Fast Stochastic Greedy} algorithm that achieves strong bicriteria approximation while substantially lowering query complexity compared to state-of-the-art methods. Our approach dramatically reduces the number of function evaluations, making it highly scalable and practical for large-scale real-world AI applications where efficiency is essential.", "date": "2025-11-02", "categories": []} {"id": "2505.13904", "title": "Learning to Insert for Constructive Neural Vehicle Routing Solver", "abstract": "Neural Combinatorial Optimisation (NCO) is a promising learning-based approach for solving Vehicle Routing Problems (VRPs) without extensive manual design. While existing constructive NCO methods typically follow an appending-based paradigm that sequentially adds unvisited nodes to partial solutions, this rigid approach often leads to suboptimal results. To overcome this limitation, we explore the idea of insertion-based paradigm and propose Learning to Construct with Insertion-based Paradigm (L2C-Insert), a novel learning-based method for constructive NCO. Unlike traditional approaches, L2C-Insert builds solutions by strategically inserting unvisited nodes at any valid position in the current partial solution, which can significantly enhance the flexibility and solution quality. The proposed framework introduces three key components: a novel model architecture for precise insertion position prediction, an efficient training scheme for model optimization, and an advanced inference technique that fully exploits the insertion paradigm's flexibility. Extensive experiments on both synthetic and real-world instances of the Travelling Salesman Problem (TSP) and Capacitated Vehicle Routing Problem (CVRP) demonstrate that L2C-Insert consistently achieves superior performance across various problem sizes.", "date": "2025-05-20", "categories": []} {"id": "2505.24627", "title": "Rethinking Neural Combinatorial Optimization for Vehicle Routing Problems with Different Constraint Tightness Degrees", "abstract": "Recent neural combinatorial optimization (NCO) methods have shown promising problem-solving ability without requiring domain-specific expertise. Most existing NCO methods use training and testing data with a fixed constraint value and lack research on the effect of constraint tightness on the performance of NCO methods. This paper takes the capacity-constrained vehicle routing problem (CVRP) as an example to empirically analyze the NCO performance under different tightness degrees of the capacity constraint. Our analysis reveals that existing NCO methods overfit the capacity constraint, and they can only perform satisfactorily on a small range of the constraint values but poorly on other values. To tackle this drawback of existing NCO methods, we develop an efficient training scheme that explicitly considers varying degrees of constraint tightness and proposes a multi-expert module to learn a generally adaptable solving strategy. Experimental results show that the proposed method can effectively overcome the overfitting issue, demonstrating superior performances on the CVRP and CVRP with time windows (CVRPTW) with various constraint tightness degrees.", "date": "2025-05-30", "categories": []} {"id": "2510.26217", "title": "Hybrid LLM and Higher-Order Quantum Approximate Optimization for CSA Collateral Management", "abstract": "We address finance-native collateral optimization under ISDA Credit Support Annexes (CSAs), where integer lots, Schedule A haircuts, RA/MTA gating, and issuer/currency/class caps create rugged, legally bounded search spaces. We introduce a certifiable hybrid pipeline purpose-built for this domain: (i) an evidence-gated LLM that extracts CSA terms to a normalized JSON (abstain-by-default, span-cited); (ii) a quantum-inspired explorer that interleaves simulated annealing with micro higher order QAOA (HO-QAOA) on binding sub-QUBOs (subset size n <= 16, order k <= 4) to coordinate multi-asset moves across caps and RA-induced discreteness; (iii) a weighted risk-aware objective (Movement, CVaR, funding-priced overshoot) with an explicit coverage window U <= Reff+B; and (iv) CP-SAT as single arbiter to certify feasibility and gaps, including a U-cap pre-check that reports the minimal feasible buffer B*. Encoding caps/rounding as higher-order terms lets HO-QAOA target the domain couplings that defeat local swaps. On government bond datasets and multi-CSA inputs, the hybrid improves a strong classical baseline (BL-3) by 9.1%, 9.6%, and 10.7% across representative harnesses, delivering better cost-movement-tail frontiers under governance settings. We release governance grade artifacts-span citations, valuation matrix audit, weight provenance, QUBO manifests, and CP-SAT traces-to make results auditable and reproducible.", "date": "2025-10-30", "categories": []} {"id": "2510.26633", "title": "Omnipresent Yet Overlooked: Heat Kernels in Combinatorial Bayesian Optimization", "abstract": "Bayesian Optimization (BO) has the potential to solve various combinatorial tasks, ranging from materials science to neural architecture search. However, BO requires specialized kernels to effectively model combinatorial domains. Recent efforts have introduced several combinatorial kernels, but the relationships among them are not well understood. To bridge this gap, we develop a unifying framework based on heat kernels, which we derive in a systematic way and express as simple closed-form expressions. Using this framework, we prove that many successful combinatorial kernels are either related or equivalent to heat kernels, and validate this theoretical claim in our experiments. Moreover, our analysis confirms and extends the results presented in Bounce: certain algorithms' performance decreases substantially when the unknown optima of the function do not have a certain structure. In contrast, heat kernels are not sensitive to the location of the optima. Lastly, we show that a fast and simple pipeline, relying on heat kernels, is able to achieve state-of-the-art results, matching or even outperforming certain slow or complex algorithms.", "date": "2025-10-30", "categories": []} {"id": "2407.11217", "title": "Faster and Simpler Greedy Algorithm for $k$-Median and $k$-Means", "abstract": "Clustering problems such as $k$-means and $k$-median are staples of unsupervised learning, and many algorithmic techniques have been developed to tackle their numerous aspects.\n In this paper, we focus on the class of greedy approximation algorithm, that attracted less attention than local-search or primal-dual counterparts. In particular, we study the recursive greedy algorithm developed by Mettu and Plaxton [SIAM J. Comp 2003]. We provide a simplification of the algorithm, allowing for faster implementation, in graph metrics or in Euclidean space, where our algorithm matches or improves the state-of-the-art.", "date": "2024-07-15", "categories": []} {"id": "2411.12571", "title": "Large Language Models for Combinatorial Optimization of Design Structure Matrix", "abstract": "Combinatorial optimization (CO) is essential for improving efficiency and performance in engineering applications. As complexity increases with larger problem sizes and more intricate dependencies, identifying the optimal solution become challenging. When it comes to real-world engineering problems, algorithms based on pure mathematical reasoning are limited and incapable to capture the contextual nuances necessary for optimization. This study explores the potential of Large Language Models (LLMs) in solving engineering CO problems by leveraging their reasoning power and contextual knowledge. We propose a novel LLM-based framework that integrates network topology and domain knowledge to optimize the sequencing of Design Structure Matrix (DSM)-a common CO problem. Our experiments on various DSM cases demonstrate that the proposed method achieves faster convergence and higher solution quality than benchmark methods. Moreover, results show that incorporating contextual domain knowledge significantly improves performance despite the choice of LLMs. These findings highlight the potential of LLMs in tackling complex real-world CO problems by combining semantic and mathematical reasoning. This approach paves the way for a new paradigm in in real-world combinatorial optimization.", "date": "2024-11-19", "categories": []} {"id": "2504.02388", "title": "Steiner Traveling Salesman Problem with Quantum Annealing", "abstract": "The Steiner Traveling Salesman Problem (STSP) is a variant of the classical Traveling Salesman Problem. The STSP involves incorporating steiner nodes, which are extra nodes not originally part of the required visit set but that can be added to the route to enhance the overall solution and minimize the total travel cost. Given the NP-hard nature of the STSP, we propose a quantum approach to address it. Specifically, we employ quantum annealing using D-Wave's hardware to explore its potential for solving this problem. To enhance computational feasibility, we develop a preprocessing method that effectively reduces the network size. Our experimental results demonstrate that this reduction technique significantly decreases the problem complexity, making the Quadratic Unconstrained Binary Optimization formulation, the standard input for quantum annealers, better suited for existing quantum hardware. Furthermore, the results highlight the potential of quantum annealing as a promising and innovative approach for solving the STSP.", "date": "2025-04-03", "categories": []} {"id": "2208.07777", "title": "Mining Large Independent Sets on Massive Graphs", "abstract": "The Maximum Independent Set problem is fundamental for extracting conflict-free structure from large graphs, with applications in scheduling, recommendation, and network analysis. However, existing heuristics can stagnate when search schedules are fixed and information from past solutions is underused, leading to wasted effort in low-quality regions of the search space. We present ARCIS, an efficient algorithm for mining large independent sets on massive graphs. ARCIS couples two main components. The first is an adaptive restart policy that refreshes exploration when progress slows. The second is Consensus-Guided Vertex Fixing, which restricts the search to the non-consensus region of the graph by fixing vertices consistently observed within a round. The consensus is maintained as a running intersection within each round, and because it is recomputed at every restart, the fixing is reversible. Vertices that later lose support are automatically unfixed and their neighborhoods re-enter the working graph, which corrects occasional mistakes while preserving progress. Experiments on 222 graphs from four benchmark suites show that ARCIS attains the best or tied-best solution quality in most instances while delivering competitive runtime and low variability. Ablation studies isolate the impact of each component, indicating that ARCIS is a practical and robust method for large-scale graph mining.", "date": "2022-08-16", "categories": []} {"id": "2505.19053", "title": "Structured Reinforcement Learning for Combinatorial Decision-Making", "abstract": "Reinforcement learning (RL) is increasingly applied to real-world problems involving complex and structured decisions, such as routing, scheduling, and assortment planning. These settings challenge standard RL algorithms, which struggle to scale, generalize, and exploit structure in the presence of combinatorial action spaces. We propose Structured Reinforcement Learning (SRL), a novel actor-critic paradigm that embeds combinatorial optimization-layers into the actor neural network. We enable end-to-end learning of the actor via Fenchel-Young losses and provide a geometric interpretation of SRL as a primal-dual algorithm in the dual of the moment polytope. Across six environments with exogenous and endogenous uncertainty, SRL matches or surpasses the performance of unstructured RL and imitation learning on static tasks and improves over these baselines by up to 92% on dynamic problems, with improved stability and convergence speed.", "date": "2025-05-25", "categories": []} {"id": "2508.21730", "title": "Freeze and Conquer: Reusable Ansatz for Solving the Traveling Salesman Problem", "abstract": "In this paper we present a variational algorithm for the Traveling Salesman Problem (TSP) that combines (i) a compact encoding of permutations, which reduces the qubit requirement too, (ii) an optimize-freeze-reuse strategy: where the circuit topology (``Ansatz'') is first optimized on a training instance by Simulated Annealing (SA), then ``frozen'' and re-used on novel instances, limited to a rapid re-optimization of only the circuit parameters. This pipeline eliminates costly structural research in testing, making the procedure immediately implementable on NISQ hardware.\n On a set of $40$ randomly generated symmetric instances that span $4 - 7$ cities, the resulting Ansatz achieves an average optimal trip sampling probability of $100\\%$ for 4 city cases, $90\\%$ for 5 city cases and $80\\%$ for 6 city cases. With 7 cities the success rate drops markedly to an average of $\\sim 20\\%$, revealing the onset of scalability limitations of the proposed method.\n The results show robust generalization ability for moderate problem sizes and indicate how freezing the Ansatz can dramatically reduce time-to-solution without degrading solution quality. The paper also discusses scalability limitations, the impact of ``warm-start'' initialization of parameters, and prospects for extension to more complex problems, such as Vehicle Routing and Job-Shop Scheduling.", "date": "2025-08-29", "categories": []} {"id": "2510.23901", "title": "RS-ORT: A Reduced-Space Branch-and-Bound Algorithm for Optimal Regression Trees", "abstract": "Mixed-integer programming (MIP) has emerged as a powerful framework for learning optimal decision trees. Yet, existing MIP approaches for regression tasks are either limited to purely binary features or become computationally intractable when continuous, large-scale data are involved. Naively binarizing continuous features sacrifices global optimality and often yields needlessly deep trees. We recast the optimal regression-tree training as a two-stage optimization problem and propose Reduced-Space Optimal Regression Trees (RS-ORT) - a specialized branch-and-bound (BB) algorithm that branches exclusively on tree-structural variables. This design guarantees the algorithm's convergence and its independence from the number of training samples. Leveraging the model's structure, we introduce several bound tightening techniques - closed-form leaf prediction, empirical threshold discretization, and exact depth-1 subtree parsing - that combine with decomposable upper and lower bounding strategies to accelerate the training. The BB node-wise decomposition enables trivial parallel execution, further alleviating the computational intractability even for million-size datasets. Based on the empirical studies on several regression benchmarks containing both binary and continuous features, RS-ORT also delivers superior training and testing performance than state-of-the-art methods. Notably, on datasets with up to 2,000,000 samples with continuous features, RS-ORT can obtain guaranteed training performance with a simpler tree structure and a better generalization ability in four hours.", "date": "2025-10-27", "categories": []} {"id": "2510.24013", "title": "Discovering Heuristics with Large Language Models (LLMs) for Mixed-Integer Programs: Single-Machine Scheduling", "abstract": "Our study contributes to the scheduling and combinatorial optimization literature with new heuristics discovered by leveraging the power of Large Language Models (LLMs). We focus on the single-machine total tardiness (SMTT) problem, which aims to minimize total tardiness by sequencing n jobs on a single processor without preemption, given processing times and due dates. We develop and benchmark two novel LLM-discovered heuristics, the EDD Challenger (EDDC) and MDD Challenger (MDDC), inspired by the well-known Earliest Due Date (EDD) and Modified Due Date (MDD) rules. In contrast to prior studies that employed simpler rule-based heuristics, we evaluate our LLM-discovered algorithms using rigorous criteria, including optimality gaps and solution time derived from a mixed-integer programming (MIP) formulation of SMTT. We compare their performance against state-of-the-art heuristics and exact methods across various job sizes (20, 100, 200, and 500 jobs). For instances with more than 100 jobs, exact methods such as MIP and dynamic programming become computationally intractable. Up to 500 jobs, EDDC improves upon the classic EDD rule and another widely used algorithm in the literature. MDDC consistently outperforms traditional heuristics and remains competitive with exact approaches, particularly on larger and more complex instances. This study shows that human-LLM collaboration can produce scalable, high-performing heuristics for NP-hard constrained combinatorial optimization, even under limited resources when effectively configured.", "date": "2025-10-28", "categories": []} {"id": "2506.12932", "title": "Complexity Scaling Laws for Neural Models using Combinatorial Optimization", "abstract": "Recent work on neural scaling laws demonstrates that model performance scales predictably with compute budget, model size, and dataset size. In this work, we develop scaling laws based on problem complexity. We analyze two fundamental complexity measures: solution space size and representation space size. Using the Traveling Salesman Problem (TSP) as a case study, we show that combinatorial optimization promotes smooth cost trends, and therefore meaningful scaling laws can be obtained even in the absence of an interpretable loss. We then show that suboptimality grows predictably for fixed-size models when scaling the number of TSP nodes or spatial dimensions, independent of whether the model was trained with reinforcement learning or supervised fine-tuning on a static dataset. We conclude with an analogy to problem complexity scaling in local search, showing that a much simpler gradient descent of the cost landscape produces similar trends.", "date": "2025-06-15", "categories": []} {"id": "2510.21886", "title": "Exploration through Generation: Applying GFlowNets to Structured Search", "abstract": "This work applies Generative Flow Networks (GFlowNets) to three graph optimization problems: the Traveling Salesperson Problem, Minimum Spanning Tree, and Shortest Path. GFlowNets are generative models that learn to sample solutions proportionally to a reward function. The models are trained using the Trajectory Balance loss to build solutions sequentially, selecting edges for spanning trees, nodes for paths, and cities for tours. Experiments on benchmark instances of varying sizes show that GFlowNets learn to find optimal solutions. For each problem type, multiple graph configurations with different numbers of nodes were tested. The generated solutions match those from classical algorithms (Dijkstra for shortest path, Kruskal for spanning trees, and exact solvers for TSP). Training convergence depends on problem complexity, with the number of episodes required for loss stabilization increasing as graph size grows. Once training converges, the generated solutions match known optima from classical algorithms across the tested instances. This work demonstrates that generative models can solve combinatorial optimization problems through learned policies. The main advantage of this learning-based approach is computational scalability: while classical algorithms have fixed complexity per instance, GFlowNets amortize computation through training. With sufficient computational resources, the framework could potentially scale to larger problem instances where classical exact methods become infeasible.", "date": "2025-10-23", "categories": []} {"id": "2510.22131", "title": "Probing Neural Combinatorial Optimization Models", "abstract": "Neural combinatorial optimization (NCO) has achieved remarkable performance, yet its learned model representations and decision rationale remain a black box. This impedes both academic research and practical deployment, since researchers and stakeholders require deeper insights into NCO models. In this paper, we take the first critical step towards interpreting NCO models by investigating their representations through various probing tasks. Moreover, we introduce a novel probing tool named Coefficient Significance Probing (CS-Probing) to enable deeper analysis of NCO representations by examining the coefficients and statistical significance during probing. Extensive experiments and analysis reveal that NCO models encode low-level information essential for solution construction, while capturing high-level knowledge to facilitate better decisions. Using CS-Probing, we find that prevalent NCO models impose varying inductive biases on their learned representations, uncover direct evidence related to model generalization, and identify key embedding dimensions associated with specific knowledge. These insights can be potentially translated into practice, for example, with minor code modifications, we improve the generalization of the analyzed model. Our work represents a first systematic attempt to interpret black-box NCO models, showcasing probing as a promising tool for analyzing their internal mechanisms and revealing insights for the NCO community. The source code is publicly available.", "date": "2025-10-25", "categories": []} {"id": "2503.23209", "title": "A QUBO Framework for Team Formation", "abstract": "The team formation problem assumes a set of experts and a task, where each expert has a set of skills and the task requires some skills. The objective is to find a set of experts that maximizes coverage of the required skills while simultaneously minimizing the costs associated with the experts. Different definitions of cost have traditionally led to distinct problem formulations and algorithmic solutions. We introduce the unified TeamFormation formulation that captures all cost definitions for team formation problems that balance task coverage and expert cost. Specifically, we formulate three TeamFormation variants with different cost functions using quadratic unconstrained binary optimization (QUBO), and we evaluate two distinct general-purpose solution methods. We show that solutions based on the QUBO formulations of TeamFormation problems are at least as good as those produced by established baselines. Furthermore, we show that QUBO-based solutions leveraging graph neural networks can effectively learn representations of experts and skills to enable transfer learning, allowing node embeddings from one problem instance to be efficiently applied to another.", "date": "2025-03-29", "categories": []} {"id": "2510.21431", "title": "Oracle-Efficient Combinatorial Semi-Bandits", "abstract": "We study the combinatorial semi-bandit problem where an agent selects a subset of base arms and receives individual feedback. While this generalizes the classical multi-armed bandit and has broad applicability, its scalability is limited by the high cost of combinatorial optimization, requiring oracle queries at every round. To tackle this, we propose oracle-efficient frameworks that significantly reduce oracle calls while maintaining tight regret guarantees. For the worst-case linear reward setting, our algorithms achieve $\\tilde{O}(\\sqrt{T})$ regret using only $O(\\log\\log T)$ oracle queries. We also propose covariance-adaptive algorithms that leverage noise structure for improved regret, and extend our approach to general (non-linear) rewards. Overall, our methods reduce oracle usage from linear to (doubly) logarithmic in time, with strong theoretical guarantees.", "date": "2025-10-24", "categories": []} {"id": "2510.21453", "title": "Multi-Task Vehicle Routing Solver via Mixture of Specialized Experts under State-Decomposable MDP", "abstract": "Existing neural methods for multi-task vehicle routing problems (VRPs) typically learn unified solvers to handle multiple constraints simultaneously. However, they often underutilize the compositional structure of VRP variants, each derivable from a common set of basis VRP variants. This critical oversight causes unified solvers to miss out the potential benefits of basis solvers, each specialized for a basis VRP variant. To overcome this limitation, we propose a framework that enables unified solvers to perceive the shared-component nature across VRP variants by proactively reusing basis solvers, while mitigating the exponential growth of trained neural solvers. Specifically, we introduce a State-Decomposable MDP (SDMDP) that reformulates VRPs by expressing the state space as the Cartesian product of basis state spaces associated with basis VRP variants. More crucially, this formulation inherently yields the optimal basis policy for each basis VRP variant. Furthermore, a Latent Space-based SDMDP extension is developed by incorporating both the optimal basis policies and a learnable mixture function to enable the policy reuse in the latent space. Under mild assumptions, this extension provably recovers the optimal unified policy of SDMDP through the mixture function that computes the state embedding as a mapping from the basis state embeddings generated by optimal basis policies. For practical implementation, we introduce the Mixture-of-Specialized-Experts Solver (MoSES), which realizes basis policies through specialized Low-Rank Adaptation (LoRA) experts, and implements the mixture function via an adaptive gating mechanism. Extensive experiments conducted across VRP variants showcase the superiority of MoSES over prior methods.", "date": "2025-10-24", "categories": []} {"id": "2405.14090", "title": "Solving 0-1 Integer Programs with Unknown Knapsack Constraints Using Membership Oracles", "abstract": "We consider solving a combinatorial optimization problem with unknown knapsack constraints using a membership oracle for each unknown constraint such that, given a solution, the oracle determines whether the constraint is satisfied or not with absolute certainty. The goal of the decision maker is to find the best possible solution subject to a budget on the number of oracle calls. Inspired by active learning for binary classification based on Support Vector Machines (SVMs), we devise a framework to solve the problem by learning and exploiting surrogate linear constraints. The framework includes training linear separators on the labeled points and selecting new points to be labeled, which is achieved by applying a sampling strategy and solving a 0-1 integer linear program. Following the active learning literature, a natural choice would be SVM as a linear classifier and the information-based sampling strategy known as simple margin, for each unknown constraint. We improve on both sides: we propose an alternative sampling strategy based on mixed-integer quadratic programming and a linear separation method inspired by an algorithm for convex optimization in the oracle model. We conduct experiments on classical problems and variants inspired by realistic applications to show how different linear separation methods and sampling strategies influence the quality of the results in terms of several metrics including objective value, dual bound and running time.", "date": "2024-05-23", "categories": []} {"id": "2503.01455", "title": "Proper decision trees: An axiomatic framework for solving optimal decision tree problems with arbitrary splitting rules", "abstract": "We present an axiomatic framework for analyzing the algorithmic properties of decision trees. This framework supports the classification of decision tree problems through structural and ancestral constraints within a rigorous mathematical foundation. The central focus of this paper is a special class of decision tree problems-which we term proper decision trees-due to their versatility and effectiveness. In terms of versatility, this class subsumes several well-known data structures, including binary space partitioning trees, K-D trees, and machine learning decision tree models. Regarding effectiveness, we prove that only proper decision trees can be uniquely characterized as K-permutations, whereas typical non-proper decision trees correspond to binary-labeled decision trees with substantially greater complexity. Using this formal characterization, we develop a generic algorithmic approach for solving optimal decision tree problems over arbitrary splitting rules and objective functions for proper decision trees. We constructively derive a generic dynamic programming recursion for solving these problems exactly. However, we show that memoization is generally impractical in terms of space complexity, as both datasets and subtrees must be stored. This result contradicts claims in the literature that suggest a trade-off between memoizing datasets and subtrees. Our framework further accommodates constraints such as tree depth and leaf size, and can be accelerated using techniques such as thinning. Finally, we extend our analysis to several non-proper decision trees, including the commonly studied decision tree over binary feature data, the binary search tree, and the tree structure arising in the matrix chain multiplication problem. We demonstrate how these problems can be solved by appropriately modifying or discarding certain axioms.", "date": "2025-03-03", "categories": []} {"id": "2507.13703", "title": "Binarizing Physics-Inspired GNNs for Combinatorial Optimization", "abstract": "Physics-inspired graph neural networks (PI-GNNs) have been utilized as an efficient unsupervised framework for relaxing combinatorial optimization problems encoded through a specific graph structure and loss, reflecting dependencies between the problem's variables. While the framework has yielded promising results in various combinatorial problems, we show that the performance of PI-GNNs systematically plummets with an increasing density of the combinatorial problem graphs. Our analysis reveals an interesting phase transition in the PI-GNNs' training dynamics, associated with degenerate solutions for the denser problems, highlighting a discrepancy between the relaxed, real-valued model outputs and the binary-valued problem solutions. To address the discrepancy, we propose principled alternatives to the naive strategy used in PI-GNNs by building on insights from fuzzy logic and binarized neural networks. Our experiments demonstrate that the portfolio of proposed methods significantly improves the performance of PI-GNNs in increasingly dense settings.", "date": "2025-07-18", "categories": []} {"id": "2510.19835", "title": "A Quantum-Inspired Algorithm for Solving Sudoku Puzzles and the MaxCut Problem", "abstract": "We propose and evaluate a quantum-inspired algorithm for solving Quadratic Unconstrained Binary Optimization (QUBO) problems, which are mathematically equivalent to finding ground states of Ising spin-glass Hamiltonians. The algorithm employs Matrix Product States (MPS) to compactly represent large superpositions of spin configurations and utilizes a discrete driving schedule to guide the MPS toward the ground state. At each step, a driver Hamiltonian -- incorporating a transverse magnetic field -- is combined with the problem Hamiltonian to enable spin flips and facilitate quantum tunneling. The MPS is updated using the standard Density Matrix Renormalization Group (DMRG) method, which iteratively minimizes the system's energy via multiple sweeps across the spin chain. Despite its heuristic nature, the algorithm reliably identifies global minima, not merely near-optimal solutions, across diverse QUBO instances. We first demonstrate its effectiveness on intermediate-level Sudoku puzzles from publicly available sources, involving over $200$ Ising spins with long-range couplings dictated by constraint satisfaction. We then apply the algorithm to MaxCut problems from the Biq Mac library, successfully solving instances with up to $251$ nodes and $3,265$ edges. We discuss the advantages of this quantum-inspired approach, including its scalability, generalizability, and suitability for industrial-scale QUBO applications.", "date": "2025-10-10", "categories": []} {"id": "2510.20431", "title": "Partial Optimality in Cubic Correlation Clustering for General Graphs", "abstract": "The higher-order correlation clustering problem for a graph $G$ and costs associated with cliques of $G$ consists in finding a clustering of $G$ so as to minimize the sum of the costs of those cliques whose nodes all belong to the same cluster. To tackle this NP-hard problem in practice, local search heuristics have been proposed and studied in the context of applications. Here, we establish partial optimality conditions for cubic correlation clustering, i.e., for the special case of at most 3-cliques. We define and implement algorithms for deciding these conditions and examine their effectiveness numerically, on two data sets.", "date": "2025-10-23", "categories": []} {"id": "2409.03811", "title": "PARCO: Parallel AutoRegressive Models for Multi-Agent Combinatorial Optimization", "abstract": "Combinatorial optimization problems involving multiple agents are notoriously challenging due to their NP-hard nature and the necessity for effective agent coordination. Despite advancements in learning-based methods, existing approaches often face critical limitations, including suboptimal agent coordination, poor generalization, and high computational latency. To address these issues, we propose PARCO (Parallel AutoRegressive Combinatorial Optimization), a general reinforcement learning framework designed to construct high-quality solutions for multi-agent combinatorial tasks efficiently. To this end, PARCO integrates three key novel components: (1) transformer-based communication layers to enable effective agent collaboration during parallel solution construction, (2) a multiple pointer mechanism for low-latency, parallel agent decision-making, and (3) priority-based conflict handlers to resolve decision conflicts via learned priorities. We evaluate PARCO in multi-agent vehicle routing and scheduling problems, where our approach outperforms state-of-the-art learning methods, demonstrating strong generalization ability and remarkable computational efficiency. We make our source code publicly available to foster future research: https://github.com/ai4co/parco.", "date": "2024-09-05", "categories": []} {"id": "2510.19348", "title": "A Markov Decision Process for Variable Selection in Branch & Bound", "abstract": "Mixed-Integer Linear Programming (MILP) is a powerful framework used to address a wide range of NP-hard combinatorial optimization problems, often solved by Branch and Bound (B&B). A key factor influencing the performance of B&B solvers is the variable selection heuristic governing branching decisions. Recent contributions have sought to adapt reinforcement learning (RL) algorithms to the B&B setting to learn optimal branching policies, through Markov Decision Processes (MDP) inspired formulations, and ad hoc convergence theorems and algorithms. In this work, we introduce BBMDP, a principled vanilla MDP formulation for variable selection in B&B, allowing to leverage a broad range of RL algorithms for the purpose of learning optimal B\\&B heuristics. Computational experiments validate our model empirically, as our branching agent outperforms prior state-of-the-art RL agents on four standard MILP benchmarks.", "date": "2025-10-22", "categories": []} {"id": "2405.18754", "title": "GIST: Greedy Independent Set Thresholding for Max-Min Diversification with Submodular Utility", "abstract": "This work studies a novel subset selection problem called max-min diversification with monotone submodular utility ($\\textsf{MDMS}$), which has a wide range of applications in machine learning, e.g., data sampling and feature selection. Given a set of points in a metric space, the goal of $\\textsf{MDMS}$ is to maximize $f(S) = g(S) + \\lambda \\cdot \\texttt{div}(S)$ subject to a cardinality constraint $|S| \\le k$, where $g(S)$ is a monotone submodular function and $\\texttt{div}(S) = \\min_{u,v \\in S : u \\ne v} \\text{dist}(u,v)$ is the max-min diversity objective. We propose the $\\texttt{GIST}$ algorithm, which gives a $\\frac{1}{2}$-approximation guarantee for $\\textsf{MDMS}$ by approximating a series of maximum independent set problems with a bicriteria greedy algorithm. We also prove that it is NP-hard to approximate within a factor of $0.5584$. Finally, we show in our empirical study that $\\texttt{GIST}$ outperforms state-of-the-art benchmarks for a single-shot data sampling task on ImageNet.", "date": "2024-05-29", "categories": []} {"id": "2510.10262", "title": "Enhancing the Cross-Size Generalization for Solving Vehicle Routing Problems via Continual Learning", "abstract": "Exploring machine learning techniques for addressing vehicle routing problems has attracted considerable research attention. To achieve decent and efficient solutions, existing deep models for vehicle routing problems are typically trained and evaluated using instances of a single size. This substantially limits their ability to generalize across different problem sizes and thus hampers their practical applicability. To address the issue, we propose a continual learning based framework that sequentially trains a deep model with instances of ascending problem sizes. Specifically, on the one hand, we design an inter-task regularization scheme to retain the knowledge acquired from smaller problem sizes in the model training on a larger size. On the other hand, we introduce an intra-task regularization scheme to consolidate the model by imitating the latest desirable behaviors during training on each size. Additionally, we exploit the experience replay to revisit instances of formerly trained sizes for mitigating the catastrophic forgetting. Experimental results show that our approach achieves predominantly superior performance across various problem sizes (either seen or unseen in the training), as compared to state-of-the-art deep models including the ones specialized for generalizability enhancement. Meanwhile, the ablation studies on the key designs manifest their synergistic effect in the proposed framework.", "date": "2025-10-11", "categories": []} {"id": "2510.16047", "title": "Algorithms for dynamic scheduling in manufacturing, towards digital factories Improving Deadline Feasibility and Responsiveness via Temporal Networks", "abstract": "Modern manufacturing systems must meet hard delivery deadlines while coping with stochastic task durations caused by process noise, equipment variability, and human intervention. Traditional deterministic schedules break down when reality deviates from nominal plans, triggering costly last-minute repairs. This thesis combines offline constraint-programming (CP) optimisation with online temporal-network execution to create schedules that remain feasible under worst-case uncertainty. First, we build a CP model of the flexible job-shop with per-job deadline tasks and insert an optimal buffer $\\Delta^*$ to obtain a fully pro-active baseline. We then translate the resulting plan into a Simple Temporal Network with Uncertainty (STNU) and verify dynamic controllability, which guarantees that a real-time dispatcher can retime activities for every bounded duration realisation without violating resource or deadline constraints. Extensive Monte-Carlo simulations on the open Kacem~1--4 benchmark suite show that our hybrid approach eliminates 100\\% of deadline violations observed in state-of-the-art meta-heuristic schedules, while adding only 3--5\\% makespan overhead. Scalability experiments confirm that CP solve-times and STNU checks remain sub-second on medium-size instances. The work demonstrates how temporal-network reasoning can bridge the gap between proactive buffering and dynamic robustness, moving industry a step closer to truly digital, self-correcting factories.", "date": "2025-10-16", "categories": []} {"id": "2510.17382", "title": "Graph Attention-Guided Search for Dense Multi-Agent Pathfinding", "abstract": "Finding near-optimal solutions for dense multi-agent pathfinding (MAPF) problems in real-time remains challenging even for state-of-the-art planners. To this end, we develop a hybrid framework that integrates a learned heuristic derived from MAGAT, a neural MAPF policy with a graph attention scheme, into a leading search-based algorithm, LaCAM. While prior work has explored learning-guided search in MAPF, such methods have historically underperformed. In contrast, our approach, termed LaGAT, outperforms both purely search-based and purely learning-based methods in dense scenarios. This is achieved through an enhanced MAGAT architecture, a pre-train-then-fine-tune strategy on maps of interest, and a deadlock detection scheme to account for imperfect neural guidance. Our results demonstrate that, when carefully designed, hybrid search offers a powerful solution for tightly coupled, challenging multi-agent coordination problems.", "date": "2025-10-20", "categories": []} {"id": "2510.15257", "title": "Minimisation of Submodular Functions Using Gaussian Zeroth-Order Random Oracles", "abstract": "We consider the minimisation problem of submodular functions and investigate the application of a zeroth-order method to this problem. The method is based on exploiting a Gaussian smoothing random oracle to estimate the smoothed function gradient. We prove the convergence of the algorithm to a global $\\epsilon$-approximate solution in the offline case and show that the algorithm is Hannan-consistent in the online case with respect to static regret. Moreover, we show that the algorithm achieves $O(\\sqrt{NP_N^\\ast})$ dynamic regret, where $N$ is the number of iterations and $P_N^\\ast$ is the path length. The complexity analysis and hyperparameter selection are presented for all the cases. The theoretical results are illustrated via numerical examples.", "date": "2025-10-17", "categories": []} {"id": "2510.14317", "title": "Column Generation Using Domain-Independent Dynamic Programming", "abstract": "Column generation and branch-and-price are leading methods for large-scale exact optimization. Column generation iterates between solving a master problem and a pricing problem. The master problem is a linear program, which can be solved using a generic solver. The pricing problem is highly dependent on the application but is usually discrete. Due to the difficulty of discrete optimization, high-performance column generation often relies on a custom pricing algorithm built specifically to exploit the problem's structure. This bespoke nature of the pricing solver prevents the reuse of components for other applications. We show that domain-independent dynamic programming, a software package for modeling and solving arbitrary dynamic programs, can be used as a generic pricing solver. We develop basic implementations of branch-and-price with pricing by domain-independent dynamic programming and show that they outperform a world-leading solver on static mixed integer programming formulations for seven problem classes.", "date": "2025-10-16", "categories": []} {"id": "2504.04821", "title": "A Customized SAT-based Solver for Graph Coloring", "abstract": "We introduce ZykovColor, a novel SAT-based algorithm to solve the graph coloring problem working on top of an encoding that mimics the Zykov tree. Our method is based on an approach of H\\'ebrard and Katsirelos (2020) that employs a propagator to enforce transitivity constraints, incorporate lower bounds for search tree pruning, and enable inferred propagations.\n We leverage the recently introduced IPASIR-UP interface for CaDiCaL to implement these techniques with a SAT solver. Furthermore, we propose new features that take advantage of the underlying SAT solver. These include modifying the integrated decision strategy with vertex domination hints and using incremental bottom-up search that allows to reuse learned clauses from previous calls. Additionally, we integrate a more effective clique computation and an algorithm for computing the fractional chromatic number to improve the lower bounds used for pruning during the search.\n We validate the effectiveness of each new feature through an experimental analysis. ZykovColor outperforms other state-of-the-art graph coloring implementations on the DIMACS benchmark set. Further experiments on random Erd\\H{o}s-R\\'enyi graphs show that our new approach matches or outperforms state-of-the-art SAT-based methods for both very sparse and highly dense graphs. We give an additional configuration of ZykovColor that dominates other SAT-based methods on the Erd\\H{o}s-R\\'enyi graphs.", "date": "2025-04-07", "categories": []} {"id": "2510.12273", "title": "Multi-Action Self-Improvement for Neural Combinatorial Optimization", "abstract": "Self-improvement has emerged as a state-of-the-art paradigm in Neural Combinatorial Optimization (NCO), where models iteratively refine their policies by generating and imitating high-quality solutions. Despite strong empirical performance, existing methods face key limitations. Training is computationally expensive, as policy updates require sampling numerous candidate solutions per instance to extract a single expert trajectory. More fundamentally, these approaches fail to exploit the structure of combinatorial problems involving the coordination of multiple agents, such as vehicles in min-max routing or machines in scheduling. By supervising on single-action trajectories, they fail to exploit agent-permutation symmetries, where distinct sequences of actions yield identical solutions, hindering generalization and the ability to learn coordinated behavior.\n We address these challenges by extending self-improvement to operate over joint multi-agent actions. Our model architecture predicts complete agent-task assignments jointly at each decision step. To explicitly leverage symmetries, we employ a set-prediction loss, which supervises the policy on multiple expert assignments for any given state. This approach enhances sample efficiency and the model's ability to learn coordinated behavior. Furthermore, by generating multi-agent actions in parallel, it drastically accelerates the solution generation phase of the self-improvement loop. Empirically, we validate our method on several combinatorial problems, demonstrating consistent improvements in the quality of the final solution and a reduced generation latency compared to standard self-improvement.", "date": "2025-10-14", "categories": []} {"id": "2510.07276", "title": "Multi-Objective Multi-Agent Path Finding with Lexicographic Cost Preferences", "abstract": "Many real-world scenarios require multiple agents to coordinate in shared environments, while balancing trade-offs between multiple, potentially competing objectives. Current multi-objective multi-agent path finding (MO-MAPF) algorithms typically produce conflict-free plans by computing Pareto frontiers. They do not explicitly optimize for user-defined preferences, even when the preferences are available, and scale poorly with the number of objectives. We propose a lexicographic framework for modeling MO-MAPF, along with an algorithm \\textit{Lexicographic Conflict-Based Search} (LCBS) that directly computes a single solution aligned with a lexicographic preference over objectives. LCBS integrates a priority-aware low-level $A^*$ search with conflict-based search, avoiding Pareto frontier construction and enabling efficient planning guided by preference over objectives. We provide insights into optimality and scalability, and empirically demonstrate that LCBS computes optimal solutions while scaling to instances with up to ten objectives -- far beyond the limits of existing MO-MAPF methods. Evaluations on standard and randomized MAPF benchmarks show consistently higher success rates against state-of-the-art baselines, especially with increasing number of objectives.", "date": "2025-10-08", "categories": []} {"id": "2510.10057", "title": "One4Many-StablePacker: An Efficient Deep Reinforcement Learning Framework for the 3D Bin Packing Problem", "abstract": "The three-dimensional bin packing problem (3D-BPP) is widely applied in logistics and warehousing. Existing learning-based approaches often neglect practical stability-related constraints and exhibit limitations in generalizing across diverse bin dimensions. To address these limitations, we propose a novel deep reinforcement learning framework, One4Many-StablePacker (O4M-SP). The primary advantage of O4M-SP is its ability to handle various bin dimensions in a single training process while incorporating support and weight constraints common in practice. Our training method introduces two innovative mechanisms. First, it employs a weighted reward function that integrates loading rate and a new height difference metric for packing layouts, promoting improved bin utilization through flatter packing configurations. Second, it combines clipped policy gradient optimization with a tailored policy drifting method to mitigate policy entropy collapse, encouraging exploration at critical decision nodes during packing to avoid suboptimal solutions. Extensive experiments demonstrate that O4M-SP generalizes successfully across diverse bin dimensions and significantly outperforms baseline methods. Furthermore, O4M-SP exhibits strong practical applicability by effectively addressing packing scenarios with stability constraints.", "date": "2025-10-11", "categories": []} {"id": "2510.11109", "title": "Graph Neural Network-Based Multicast Routing for On-Demand Streaming Services in 6G Networks", "abstract": "The increase of bandwidth-intensive applications in sixth-generation (6G) wireless networks, such as real-time volumetric streaming and multi-sensory extended reality, demands intelligent multicast routing solutions capable of delivering differentiated quality-of-service (QoS) at scale. Traditional shortest-path and multicast routing algorithms are either computationally prohibitive or structurally rigid, and they often fail to support heterogeneous user demands, leading to suboptimal resource utilization. Neural network-based approaches, while offering improved inference speed, typically lack topological generalization and scalability. To address these limitations, this paper presents a graph neural network (GNN)-based multicast routing framework that jointly minimizes total transmission cost and supports user-specific video quality requirements. The routing problem is formulated as a constrained minimum-flow optimization task, and a reinforcement learning algorithm is developed to sequentially construct efficient multicast trees by reusing paths and adapting to network dynamics. A graph attention network (GAT) is employed as the encoder to extract context-aware node embeddings, while a long short-term memory (LSTM) module models the sequential dependencies in routing decisions. Extensive simulations demonstrate that the proposed method closely approximates optimal dynamic programming-based solutions while significantly reducing computational complexity. The results also confirm strong generalization to large-scale and dynamic network topologies, highlighting the method's potential for real-time deployment in 6G multimedia delivery scenarios. Code is available at https://github.com/UNIC-Lab/GNN-Routing.", "date": "2025-10-13", "categories": []} {"id": "2510.11121", "title": "Refining Hybrid Genetic Search for CVRP via Reinforcement Learning-Finetuned LLM", "abstract": "While large language models (LLMs) are increasingly used as automated heuristic designers for vehicle routing problems (VRPs), current state-of-the-art methods predominantly rely on prompting massive, general-purpose models like GPT-4. This work challenges that paradigm by demonstrating that a smaller, specialized LLM, when meticulously fine-tuned, can generate components that surpass expert-crafted heuristics within advanced solvers. We propose RFTHGS, a novel Reinforcement learning (RL) framework for Fine-Tuning a small LLM to generate high-performance crossover operators for the Hybrid Genetic Search (HGS) solver, applied to the Capacitated VRP (CVRP). Our method employs a multi-tiered, curriculum-based reward function that progressively guides the LLM to master generating first compilable, then executable, and finally, superior-performing operators that exceed human expert designs. This is coupled with an operator caching mechanism that discourages plagiarism and promotes diversity during training. Comprehensive experiments show that our fine-tuned LLM produces crossover operators which significantly outperform the expert-designed ones in HGS. The performance advantage remains consistent, generalizing from small-scale instances to large-scale problems with up to 1000 nodes. Furthermore, RFTHGS exceeds the performance of leading neuro-combinatorial baselines, prompt-based methods, and commercial LLMs such as GPT-4o and GPT-4o-mini.", "date": "2025-10-13", "categories": []} {"id": "2508.01774", "title": "VAGPO: Vision-augmented Asymmetric Group Preference Optimization for Graph Routing Problems", "abstract": "Graph routing problems play a vital role in web-related networks, where finding optimal paths across graphs is essential for efficient data transmission and content delivery. Classic routing formulations such as the Traveling Salesman Problem (TSP) and the Capacitated Vehicle Routing Problem (CVRP) represent fundamental graph optimization challenges. Recent data-driven optimization methods have made significant progress, yet they often face limitations in training efficiency and generalization to large-scale instances. In this paper, we propose a novel Vision-augmented Asymmetric Group Preference Optimization (VAGPO) approach. By leveraging ResNet-based visual encoding and Transformer-based sequential modeling, VAGPO captures both spatial structure and temporal dependencies. Furthermore, we introduce an asymmetric group preference optimization strategy that significantly accelerates convergence compared to commonly used policy gradient methods. Experimental results on generated TSP and CVRP instances, as well as real-world datasets, demonstrate that the proposed VAGPO approach achieves highly competitive solution quality. Additionally, VAGPO exhibits strong generalization to larger instances (up to 1000 nodes) without re-training, highlighting its effectiveness in both learning efficiency and scalability.", "date": "2025-08-03", "categories": []} {"id": "2510.09373", "title": "Sequence Variables: A Constraint Programming Computational Domain for Routing and Sequencing", "abstract": "Constraint Programming (CP) offers an intuitive, declarative framework for modeling Vehicle Routing Problems (VRP), yet classical CP models based on successor variables cannot always deal with optional visits or insertion based heuristics. To address these limitations, this paper formalizes sequence variables within CP. Unlike the classical successor models, this computational domain handle optional visits and support insertion heuristics, including insertion-based Large Neighborhood Search. We provide a clear definition of their domain, update operations, and introduce consistency levels for constraints on this domain. An implementation is described with the underlying data structures required for integrating sequence variables into existing trail-based CP solvers. Furthermore, global constraints specifically designed for sequence variables and vehicle routing are introduced. Finally, the effectiveness of sequence variables is demonstrated by simplifying problem modeling and achieving competitive computational performance on the Dial-a-Ride Problem.", "date": "2025-10-10", "categories": []} {"id": "2310.07979", "title": "Graph-SCP: Accelerating Set Cover Problems with Graph Neural Networks", "abstract": "Machine learning (ML) approaches are increasingly being used to accelerate combinatorial optimization (CO) problems. We investigate the Set Cover Problem (SCP) and propose Graph-SCP, a graph neural network method that augments existing optimization solvers by learning to identify a smaller sub-problem that contains the solution space. Graph-SCP uses both supervised learning from prior solved instances and unsupervised learning to minimize the SCP objective. We evaluate the performance of Graph-SCP on synthetically weighted and unweighted SCP instances with diverse problem characteristics and complexities, and on instances from the OR Library, a canonical benchmark for SCP. We show that Graph-SCP reduces the problem size by 60-80% and achieves runtime speedups of up to 10x on average when compared to Gurobi (a state-of-the-art commercial solver), while maintaining solution quality. This is in contrast to fast greedy solutions that significantly compromise solution quality to achieve guaranteed polynomial runtime. We showcase Graph-SCP's ability to generalize to larger problem sizes, training on SCP instances with up to 3,000 subsets and testing on SCP instances with up to 10,000 subsets.", "date": "2023-10-12", "categories": []} {"id": "2404.01930", "title": "Adaptive Combinatorial Maximization: Beyond Approximate Greedy Policies", "abstract": "We study adaptive combinatorial maximization, which is a core challenge in machine learning, with applications in active learning as well as many other domains. We study the Bayesian setting, and consider the objectives of maximization under a cardinality constraint and minimum cost coverage. We provide new comprehensive approximation guarantees that subsume previous results, as well as considerably strengthen them. Our approximation guarantees simultaneously support the maximal gain ratio as well as near-submodular utility functions, and include both maximization under a cardinality constraint and a minimum cost coverage guarantee. In addition, we provided an approximation guarantee for a modified prior, which is crucial for obtaining active learning guarantees that do not depend on the smallest probability in the prior. Moreover, we discover a new parameter of adaptive selection policies, which we term the \"maximal gain ratio\". We show that this parameter is strictly less restrictive than the greedy approximation parameter that has been used in previous approximation guarantees, and show that it can be used to provide stronger approximation guarantees than previous results. In particular, we show that the maximal gain ratio is never larger than the greedy approximation factor of a policy, and that it can be considerably smaller. This provides a new insight into the properties that make a policy useful for adaptive combinatorial maximization.", "date": "2024-04-02", "categories": []} {"id": "2507.00218", "title": "Learning for routing: A guided review of recent developments and future directions", "abstract": "This paper reviews the current progress in applying machine learning (ML) tools to solve NP-hard combinatorial optimization problems, with a focus on routing problems such as the traveling salesman problem (TSP) and the vehicle routing problem (VRP). Due to the inherent complexity of these problems, exact algorithms often require excessive computational time to find optimal solutions, while heuristics can only provide approximate solutions without guaranteeing optimality. With the recent success of machine learning models, there is a growing trend in proposing and implementing diverse ML techniques to enhance the resolution of these challenging routing problems. We propose a taxonomy categorizing ML-based routing methods into construction-based and improvement-based approaches, highlighting their applicability to various problem characteristics. This review aims to integrate traditional OR methods with state-of-the-art ML techniques, providing a structured framework to guide future research and address emerging VRP variants.", "date": "2025-06-30", "categories": []} {"id": "2510.06517", "title": "Visualizing Multimodality in Combinatorial Search Landscapes", "abstract": "This work walks through different visualization techniques for combinatorial search landscapes, focusing on multimodality. We discuss different techniques from the landscape analysis literature, and how they can be combined to provide a more comprehensive view of the search landscape. We also include examples and discuss relevant work to show how others have used these techniques in practice, based on the geometric and aesthetic elements of the Grammar of Graphics. We conclude that there is no free lunch in visualization, and provide recommendations for future work as there are several paths to continue the work in this field.", "date": "2025-10-07", "categories": []} {"id": "2510.07073", "title": "VRPAgent: LLM-Driven Discovery of Heuristic Operators for Vehicle Routing Problems", "abstract": "Designing high-performing heuristics for vehicle routing problems (VRPs) is a complex task that requires both intuition and deep domain knowledge. Large language model (LLM)-based code generation has recently shown promise across many domains, but it still falls short of producing heuristics that rival those crafted by human experts. In this paper, we propose VRPAgent, a framework that integrates LLM-generated components into a metaheuristic and refines them through a novel genetic search. By using the LLM to generate problem-specific operators, embedded within a generic metaheuristic framework, VRPAgent keeps tasks manageable, guarantees correctness, and still enables the discovery of novel and powerful strategies. Across multiple problems, including the capacitated VRP, the VRP with time windows, and the prize-collecting VRP, our method discovers heuristic operators that outperform handcrafted methods and recent learning-based approaches while requiring only a single CPU core. To our knowledge, \\VRPAgent is the first LLM-based paradigm to advance the state-of-the-art in VRPs, highlighting a promising future for automated heuristics discovery.", "date": "2025-10-08", "categories": []} {"id": "2502.08397", "title": "Strong bounds for large-scale Minimum Sum-of-Squares Clustering", "abstract": "Clustering is a fundamental technique in data analysis and machine learning, used to group similar data points together. Among various clustering methods, the Minimum Sum-of-Squares Clustering (MSSC) is one of the most widely used. MSSC aims to minimize the total squared Euclidean distance between data points and their corresponding cluster centroids. Due to the unsupervised nature of clustering, achieving global optimality is crucial, yet computationally challenging. The complexity of finding the global solution increases exponentially with the number of data points, making exact methods impractical for large-scale datasets. Even obtaining strong lower bounds on the optimal MSSC objective value is computationally prohibitive, making it difficult to assess the quality of heuristic solutions. We address this challenge by introducing a novel method to validate heuristic MSSC solutions through optimality gaps. Our approach employs a divide-and-conquer strategy, decomposing the problem into smaller instances that can be handled by an exact solver. The decomposition is guided by an auxiliary optimization problem, the \"anticlustering problem\", for which we design an efficient heuristic. Computational experiments demonstrate the effectiveness of the method for large-scale instances, achieving optimality gaps below 3% in most cases while maintaining reasonable computational times. These results highlight the practicality of our approach in assessing feasible clustering solutions for large datasets, bridging a critical gap in MSSC evaluation.", "date": "2025-02-12", "categories": []} {"id": "2508.08477", "title": "A Fast GRASP Metaheuristic for the Trigger Arc TSP with MIP-Based Construction and Multi-Neighborhood Local Search", "abstract": "The Trigger Arc Traveling Salesman Problem (TA-TSP) extends the classical TSP by introducing dynamic arc costs that change when specific \"trigger\" arcs are traversed, modeling scenarios such as warehouse operations with compactable storage systems. This paper introduces a GRASP-based metaheuristic that combines multiple construction heuristics with a multi-neighborhood local search. The construction phase uses mixed-integer programming (MIP) techniques to transform the TA-TSP into a sequence of tailored TSP instances, while the improvement phase applies 2-Opt, Swap, and Relocate operators. Computational experiments on MESS 2024 competition instances achieved average optimality gaps of 0.77% and 0.40% relative to the best-known solutions within a 60-second limit. On smaller, synthetically generated datasets, the method produced solutions 11.3% better than the Gurobi solver under the same time constraints. The algorithm finished in the top three at MESS 2024, demonstrating its suitability for real-time routing applications with state-dependent travel costs.", "date": "2025-08-11", "categories": []} {"id": "2508.13703", "title": "Minimizing the Weighted Number of Tardy Jobs: Data-Driven Heuristic for Single-Machine Scheduling", "abstract": "Existing research on single-machine scheduling is largely focused on exact algorithms, which perform well on typical instances but can significantly deteriorate on certain regions of the problem space. In contrast, data-driven approaches provide strong and scalable performance when tailored to the structure of specific datasets. Leveraging this idea, we focus on a single-machine scheduling problem where each job is defined by its weight, duration, due date, and deadline, aiming to minimize the total weight of tardy jobs. We introduce a novel data-driven scheduling heuristic that combines machine learning with problem-specific characteristics, ensuring feasible solutions, which is a common challenge for ML-based algorithms. Experimental results demonstrate that our approach significantly outperforms the state-of-the-art in terms of optimality gap, number of optimal solutions, and adaptability across varied data scenarios, highlighting its flexibility for practical applications. In addition, we conduct a systematic exploration of ML models, addressing a common gap in similar studies by offering a detailed model selection process and providing insights into why the chosen model is the best fit.", "date": "2025-08-19", "categories": []} {"id": "2402.14048", "title": "PolyNet: Learning Diverse Solution Strategies for Neural Combinatorial Optimization", "abstract": "Reinforcement learning-based methods for constructing solutions to combinatorial optimization problems are rapidly approaching the performance of human-designed algorithms. To further narrow the gap, learning-based approaches must efficiently explore the solution space during the search process. Recent approaches artificially increase exploration by enforcing diverse solution generation through handcrafted rules, however, these rules can impair solution quality and are difficult to design for more complex problems. In this paper, we introduce PolyNet, an approach for improving exploration of the solution space by learning complementary solution strategies. In contrast to other works, PolyNet uses only a single-decoder and a training schema that does not enforce diverse solution generation through handcrafted rules. We evaluate PolyNet on four combinatorial optimization problems and observe that the implicit diversity mechanism allows PolyNet to find better solutions than approaches that explicitly enforce diverse solution generation.", "date": "2024-02-21", "categories": []} {"id": "2407.13968", "title": "Optimizing Agricultural Order Fulfillment Systems: A Hybrid Tree Search Approach", "abstract": "Efficient order fulfillment is vital in the agricultural industry, particularly due to the seasonal nature of seed supply chains. This paper addresses the challenge of optimizing seed orders fulfillment in a centralized warehouse where orders are processed in waves, taking into account the unpredictable arrival of seed stocks and strict order deadlines. We model the wave scheduling problem as a Markov decision process and propose an adaptive hybrid tree search algorithm that combines Monte Carlo tree search with domain-specific knowledge to efficiently navigate the complex, dynamic environment of seed distribution. By leveraging historical data and stochastic modeling, our method enables forecast-informed scheduling decisions that balance immediate requirements with long-term operational efficiency. The key idea is that we can augment Monte Carlo tree search algorithm with problem-specific side information that dynamically reduces the number of candidate actions at each decision step to handle the large state and action spaces that render traditional solution methods computationally intractable. Extensive simulations with realistic parameters, including a diverse range of products, a high volume of orders, and authentic seasonal durations, demonstrate that the proposed approach significantly outperforms existing industry standard methods.", "date": "2024-07-19", "categories": []} {"id": "2501.03715", "title": "Neural Deconstruction Search for Vehicle Routing Problems", "abstract": "Autoregressive construction approaches generate solutions to vehicle routing problems in a step-by-step fashion, leading to high-quality solutions that are nearing the performance achieved by handcrafted operations research techniques. In this work, we challenge the conventional paradigm of sequential solution construction and introduce an iterative search framework where solutions are instead deconstructed by a neural policy. Throughout the search, the neural policy collaborates with a simple greedy insertion algorithm to rebuild the deconstructed solutions. Our approach matches or surpasses the performance of state-of-the-art operations research methods across three challenging vehicle routing problems of various problem sizes.", "date": "2025-01-07", "categories": []} {"id": "2503.05323", "title": "Graph Alignment via Birkhoff Relaxation", "abstract": "We consider the graph alignment problem, wherein the objective is to find a vertex correspondence between two graphs that maximizes the edge overlap. The graph alignment problem is an instance of the quadratic assignment problem (QAP), known to be NP-hard in the worst case even to approximately solve. In this paper, we analyze Birkhoff relaxation, a tight convex relaxation of QAP, and present theoretical guarantees on its performance when the inputs follow the Gaussian Wigner Model. More specifically, the weighted adjacency matrices are correlated Gaussian Orthogonal Ensemble with correlation $1/\\sqrt{1+\\sigma^2}$. Denote the optimal solutions of the QAP and Birkhoff relaxation by $\\Pi^\\star$ and $X^\\star$ respectively. We show that $\\|X^\\star-\\Pi^\\star\\|_F^2 = o(n)$ when $\\sigma = o(n^{-1.25})$ and $\\|X^\\star-\\Pi^\\star\\|_F^2 = \\Omega(n)$ when $\\sigma = \\Omega(n^{-0.5})$. Thus, the optimal solution $X^\\star$ transitions from a small perturbation of $\\Pi^\\star$ for small $\\sigma$ to being well separated from $\\Pi^\\star$ as $\\sigma$ becomes larger than $n^{-0.5}$. This result allows us to guarantee that simple rounding procedures on $X^\\star$ align $1-o(1)$ fraction of vertices correctly whenever $\\sigma = o(n^{-1.25})$. This condition on $\\sigma$ to ensure the success of the Birkhoff relaxation is state-of-the-art.", "date": "2025-03-07", "categories": []} {"id": "2510.03377", "title": "Refined Iterated Pareto Greedy for Energy-aware Hybrid Flowshop Scheduling with Blocking Constraints", "abstract": "The scarcity of non-renewable energy sources, geopolitical problems in its supply, increasing prices, and the impact of climate change, force the global economy to develop more energy-efficient solutions for their operations. The Manufacturing sector is not excluded from this challenge as one of the largest consumers of energy. Energy-efficient scheduling is a method that attracts manufacturing companies to reduce their consumption as it can be quickly deployed and can show impact immediately. In this study, the hybrid flow shop scheduling problem with blocking constraint (BHFS) is investigated in which we seek to minimize the latest completion time (i.e. makespan) and overall energy consumption, a typical manufacturing setting across many industries from automotive to pharmaceutical. Energy consumption and the latest completion time of customer orders are usually conflicting objectives. Therefore, we first formulate the problem as a novel multi-objective mixed integer programming (MIP) model and propose an augmented epsilon-constraint method for finding the Pareto-optimal solutions. Also, an effective multi-objective metaheuristic algorithm. Refined Iterated Pareto Greedy (RIPG), is developed to solve large instances in reasonable time. Our proposed methods are benchmarked using small, medium, and large-size instances to evaluate their efficiency. Two well-known algorithms are adopted for comparing our novel approaches. The computational results show the effectiveness of our method.", "date": "2025-10-03", "categories": []} {"id": "2510.03438", "title": "Scalable Ground Station Selection for Large LEO Constellations", "abstract": "Effective ground station selection is critical for low Earth orbiting (LEO) satellite constellations to minimize operational costs, maximize data downlink volume, and reduce communication gaps between access windows. Traditional ground station selection typically begins by choosing from a fixed set of locations offered by Ground Station-as-a-Service (GSaaS) providers, which helps reduce the problem scope to optimizing locations over existing infrastructure. However, finding a globally optimal solution for stations using existing mixed-integer programming methods quickly becomes intractable at scale, especially when considering multiple providers and large satellite constellations. To address this issue, we introduce a scalable, hierarchical framework that decomposes the global selection problem into single-satellite, short time-window subproblems. Optimal station choices from each subproblem are clustered to identify consistently high-value locations across all decomposed cases. Cluster-level sets are then matched back to the closest GSaaS candidate sites to produce a globally feasible solution. This approach enables scalable coordination while maintaining near-optimal performance. We evaluate our method's performance on synthetic Walker-Star test cases (1-10 satellites, 1-10 stations), achieving solutions within 95% of the global IP optimum for all test cases. Real-world evaluations on Capella Space (5 satellites), ICEYE (40), and Planet's Flock (96) show that while exact IP solutions fail to scale, our framework continues to deliver high-quality site selections.", "date": "2025-10-03", "categories": []} {"id": "2510.04273", "title": "Influence branching for learning to solve mixed-integer programs online", "abstract": "On the occasion of the 20th Mixed Integer Program Workshop's computational competition, this work introduces a new approach for learning to solve MIPs online. Influence branching, a new graph-oriented variable selection strategy, is applied throughout the first iterations of the branch and bound algorithm. This branching heuristic is optimized online with Thompson sampling, which ranks the best graph representations of MIP's structure according to computational speed up over SCIP. We achieve results comparable to state of the art online learning methods. Moreover, our results indicate that our method generalizes well to more general online frameworks, where variations in constraint matrix, constraint vector and objective coefficients can all occur and where more samples are available.", "date": "2025-10-05", "categories": []} {"id": "2510.04792", "title": "Hybrid-Balance GFlowNet for Solving Vehicle Routing Problems", "abstract": "Existing GFlowNet-based methods for vehicle routing problems (VRPs) typically employ Trajectory Balance (TB) to achieve global optimization but often neglect important aspects of local optimization. While Detailed Balance (DB) addresses local optimization more effectively, it alone falls short in solving VRPs, which inherently require holistic trajectory optimization. To address these limitations, we introduce the Hybrid-Balance GFlowNet (HBG) framework, which uniquely integrates TB and DB in a principled and adaptive manner by aligning their intrinsically complementary strengths. Additionally, we propose a specialized inference strategy for depot-centric scenarios like the Capacitated Vehicle Routing Problem (CVRP), leveraging the depot node's greater flexibility in selecting successors. Despite this specialization, HBG maintains broad applicability, extending effectively to problems without explicit depots, such as the Traveling Salesman Problem (TSP). We evaluate HBG by integrating it into two established GFlowNet-based solvers, i.e., AGFN and GFACS, and demonstrate consistent and significant improvements across both CVRP and TSP, underscoring the enhanced solution quality and generalization afforded by our approach.", "date": "2025-10-06", "categories": []} {"id": "2406.08534", "title": "Optimizing Container Loading and Unloading through Dual-Cycling and Dockyard Rehandle Reduction Using a Hybrid Genetic Algorithm", "abstract": "This paper addresses the NP-hard problem of optimizing container handling at ports by integrating Quay Crane Dual-Cycling (QCDC) and dockyard rehandle minimization. We realized that there are interdependencies between the unloading sequence of QCDC and the dockyard plan and propose the Quay Crane Dual Cycle - Dockyard Rehandle Genetic Algorithm (QCDC-DR-GA), a hybrid Genetic Algorithm (GA) that holistically optimizes both aspects: maximizing the number of Dual Cycles (DCs) and minimizing the number of dockyard rehandles. QCDC-DR-GA employs specialized crossover and mutation strategies. Extensive experiments on various ship sizes demonstrate that QCDC-DR-GA reduces total operation time by 15-20% for large ships compared to existing methods. Statistical validation via two-tailed paired t-tests confirms significant improvements at a 5% significance level. The results underscore the inefficiency of isolated optimization and highlight the critical need for integrated algorithms in port operations. This approach increases resource utilization and operational efficiency, offering a cost-effective solution for ports to decrease turnaround times without infrastructure investments.", "date": "2024-06-12", "categories": []} {"id": "2510.03065", "title": "A Unified Deep Reinforcement Learning Approach for Close Enough Traveling Salesman Problem", "abstract": "In recent years, deep reinforcement learning (DRL) has gained traction for solving the NP-hard traveling salesman problem (TSP). However, limited attention has been given to the close-enough TSP (CETSP), primarily due to the challenge introduced by its neighborhood-based visitation criterion, wherein a node is considered visited if the agent enters a compact neighborhood around it. In this work, we formulate a Markov decision process (MDP) for CETSP using a discretization scheme and propose a novel unified dual-decoder DRL (UD3RL) framework that separates decision-making into node selection and waypoint determination. Specifically, an adapted encoder is employed for effective feature extraction, followed by a node-decoder and a loc-decoder to handle the two sub-tasks, respectively. A k-nearest neighbors subgraph interaction strategy is further introduced to enhance spatial reasoning during location decoding. Furthermore, we customize the REINFORCE algorithm to train UD3RL as a unified model capable of generalizing across different problem sizes and varying neighborhood radius types (i.e., constant and random radii). Experimental results show that UD3RL outperforms conventional methods in both solution quality and runtime, while exhibiting strong generalization across problem scales, spatial distributions, and radius ranges, as well as robustness to dynamic environments.", "date": "2025-10-03", "categories": []} {"id": "2509.22255", "title": "Evaluating LLMs for Combinatorial Optimization: One-Phase and Two-Phase Heuristics for 2D Bin-Packing", "abstract": "This paper presents an evaluation framework for assessing Large Language Models' (LLMs) capabilities in combinatorial optimization, specifically addressing the 2D bin-packing problem. We introduce a systematic methodology that combines LLMs with evolutionary algorithms to generate and refine heuristic solutions iteratively. Through comprehensive experiments comparing LLM generated heuristics against traditional approaches (Finite First-Fit and Hybrid First-Fit), we demonstrate that LLMs can produce more efficient solutions while requiring fewer computational resources. Our evaluation reveals that GPT-4o achieves optimal solutions within two iterations, reducing average bin usage from 16 to 15 bins while improving space utilization from 0.76-0.78 to 0.83. This work contributes to understanding LLM evaluation in specialized domains and establishes benchmarks for assessing LLM performance in combinatorial optimization tasks.", "date": "2025-09-26", "categories": []} {"id": "2510.01693", "title": "PASTA: A Unified Framework for Offline Assortment Learning", "abstract": "We study a broad class of assortment optimization problems in an offline and data-driven setting. In such problems, a firm lacks prior knowledge of the underlying choice model, and aims to determine an optimal assortment based on historical customer choice data. The combinatorial nature of assortment optimization often results in insufficient data coverage, posing a significant challenge in designing provably effective solutions. To address this, we introduce a novel Pessimistic Assortment Optimization (PASTA) framework that leverages the principle of pessimism to achieve optimal expected revenue under general choice models. Notably, PASTA requires only that the offline data distribution contains an optimal assortment, rather than providing the full coverage of all feasible assortments. Theoretically, we establish the first finite-sample regret bounds for offline assortment optimization across several widely used choice models, including the multinomial logit and nested logit models. Additionally, we derive a minimax regret lower bound, proving that PASTA is minimax optimal in terms of sample and model complexity. Numerical experiments further demonstrate that our method outperforms existing baseline approaches.", "date": "2025-10-02", "categories": []} {"id": "2510.00958", "title": "Test-Time Search in Neural Graph Coarsening Procedures for the Capacitated Vehicle Routing Problem", "abstract": "The identification of valid inequalities, such as the rounded capacity inequalities (RCIs), is a key component of cutting plane methods for the Capacitated Vehicle Routing Problem (CVRP). While a deep learning-based separation method can learn to find high-quality cuts, our analysis reveals that the model produces fewer cuts than expected because it is insufficiently sensitive to generate a diverse set of generated subsets. This paper proposes an alternative: enhancing the performance of a trained model at inference time through a new test-time search with stochasticity. First, we introduce stochastic edge selection into the graph coarsening procedure, replacing the previously proposed greedy approach. Second, we propose the Graph Coarsening History-based Partitioning (GraphCHiP) algorithm, which leverages coarsening history to identify not only RCIs but also, for the first time, the Framed capacity inequalities (FCIs). Experiments on randomly generated CVRP instances demonstrate the effectiveness of our approach in reducing the dual gap compared to the existing neural separation method. Additionally, our method discovers effective FCIs on a specific instance, despite the challenging nature of identifying such cuts.", "date": "2025-10-01", "categories": []} {"id": "2510.01094", "title": "Optimizing Fairness in Production Planning: A Human-Centric Approach to Machine and Workforce Allocation", "abstract": "This work presents a two-layer, human-centric production planning framework designed to optimize both operational efficiency and workforce fairness in industrial manufacturing. The first layer formulates the Order-Line allocation as a Constraint Programming (CP) problem, generating high-utilization production schedules that respect machine capacities, processing times, and due dates. The second layer models Worker-Line allocation as a Markov Decision Process (MDP), integrating human factors such as worker preference, experience, resilience, and medical constraints into the assignment process. Three solution strategies, greedy allocation, MCTS, and RL, are implemented and compared across multiple evaluation scenarios. The proposed system is validated through 16 test sessions with domain experts from the automotive industry, combining quantitative key performance indicators (KPIs) with expert ratings. Results indicate that the CP-based scheduling approach produces compact, feasible production plans with low tardiness, while the MDP-based worker allocation significantly improves fairness and preference alignment compared to baseline approaches. Domain experts rated both the Order-Line and Worker-Line components as effective and highlighted opportunities to further refine the objective function to penalize excessive earliness and improve continuity in worker assignments. Overall, the findings demonstrate that combining CP with learning-based decision-making provides a robust approach for human-centric production planning. The approach enables simultaneous optimization of throughput and workforce well-being, offering a practical foundation for fair and efficient manufacturing scheduling in industrial settings.", "date": "2025-10-01", "categories": []} {"id": "2509.26094", "title": "On Computing Top-$k$ Simple Shortest Paths from a Single Source", "abstract": "We investigate the problem of computing the top-$k$ simple shortest paths in weighted digraphs. While the single-pair variant -- finding the top-$k$ simple shortest paths between two specified vertices -- has been extensively studied over the past decades, with Yen's algorithm and its heuristic improvements emerging as the most effective solving strategies, relatively little attention has been devoted to the more general single-source version, where the goal is determining top-$k$ simple shortest paths from a source vertex to all other vertices. Motivated by the numerous practical applications of ranked shortest paths, in this paper we provide new insights and algorithmic contributions to this problem. In particular, we first present a theoretical characterization of the structural properties of its solutions. Then, we introduce the first polynomial-time algorithm specifically designed to handle it. On the one hand, we prove our new algorithm is on par, in terms of time complexity, with the best (and only) polynomial-time approach known in the literature to solve the problem, that is applying the fastest single-pair algorithm independently to each vertex pair formed by the source and the remaining vertices. On the other hand, through an extensive experimental evaluation on both real-world and synthetic graphs, we demonstrate that our algorithm consistently and significantly outperforms the latter baseline in terms of running time, achieving speed-ups of up to several orders of magnitude. These results establish our new algorithm as the solution to be preferred for computing $k$ simple shortest paths from a single source in practical settings.", "date": "2025-09-30", "categories": []} {"id": "2509.26216", "title": "Comparative Analysis of Ant Colony Optimization and Google OR-Tools for Solving the Open Capacitated Vehicle Routing Problem in Logistics", "abstract": "In modern logistics management systems, route planning requires high efficiency. The Open Capacitated Vehicle Routing Problem (OCVRP) deals with finding optimal delivery routes for a fleet of vehicles serving geographically distributed customers, without requiring the vehicles to return to the depot after deliveries. The present study is comparative in nature and speaks of two algorithms for OCVRP solution: Ant Colony Optimization (ACO), a nature-inspired metaheuristic; and Google OR-Tools, an industry-standard toolkit for optimization. Both implementations were developed in Python and using a custom dataset. Performance appraisal was based on routing efficiency, computation time, and scalability. The results show that ACO allows flexibility in routing parameters while OR-Tools runs much faster with more consistency and requires less input. This could help choose among routing strategies for scalable real-time logistics systems.", "date": "2025-09-30", "categories": []} {"id": "2509.26229", "title": "Hybrid Quantum-Classical Optimisation of Traveling Salesperson Problem", "abstract": "The Traveling Salesperson Problem (TSP), a quintessential NP-hard combinatorial optimisation challenge, is vital for logistics and network design but limited by exponential complexity in large instances. We propose a hybrid quantum-classical framework integrating variational quantum eigensolver (VQE) optimisation with classical machine learning, using K-means clustering for problem decomposition and a RandomForestRegressor for path refinement. Evaluated on 80 European cities (from 4 to 80 cities, 38,500 samples in total) via Qiskit's AerSimulator and ibm_kyiv 127-qubit backend, the hybrid approach outperforms quantum-only methods, achieving an approximation ratio of 1.0287 at 80 cities, a 47.5% improvement over quantum-only's 1.9614, nearing the classical baseline. Machine learning reduces variability in tour distances (interquartile range, IQR - the spread of the middle 50% of results relative to the median - from 0.06 to 0.04), enhancing stability despite noisy intermediate-scale quantum (NISQ) noise. This framework underscores hybrid strategies' potential for scalable TSP optimisation, with future hardware advancements promising practical quantum advantages.", "date": "2025-09-30", "categories": []} {"id": "2408.16717", "title": "A GREAT Architecture for Edge-Based Graph Problems Like TSP", "abstract": "In the last years, an increasing number of learning-based approaches have been proposed to tackle combinatorial optimization problems such as routing problems. Many of these approaches are based on graph neural networks (GNNs) or related transformers, operating on the Euclidean coordinates representing the routing problems. However, such models are ill-suited for a wide range of real-world problems that feature non-Euclidean and asymmetric edge costs. To overcome this limitation, we propose a novel GNN-based and edge-focused neural model called Graph Edge Attention Network (GREAT). Using GREAT as an encoder to capture the properties of a routing problem instance, we build a reinforcement learning framework which we apply to both Euclidean and non-Euclidean variants of vehicle routing problems such as Traveling Salesman Problem, Capacitated Vehicle Routing Problem and Orienteering Problem. Our framework is among the first to tackle non-Euclidean variants of these problems and achieves competitive results among learning-based benchmarks.", "date": "2024-08-29", "categories": []} {"id": "2505.14240", "title": "Learning with Local Search MCMC Layers", "abstract": "Integrating combinatorial optimization layers into neural networks has recently attracted significant research interest. However, many existing approaches lack theoretical guarantees or fail to perform adequately when relying on inexact solvers. This is a critical limitation, as many operations research problems are NP-hard, often necessitating the use of neighborhood-based local search heuristics. These heuristics iteratively generate and evaluate candidate solutions based on an acceptance rule. In this paper, we introduce a theoretically-principled approach for learning with such inexact combinatorial solvers. Inspired by the connection between simulated annealing and Metropolis-Hastings, we propose to transform problem-specific neighborhood systems used in local search heuristics into proposal distributions, implementing MCMC on the combinatorial space of feasible solutions. This allows us to construct differentiable combinatorial layers and associated loss functions. Replacing an exact solver by a local search strongly reduces the computational burden of learning on many applications. We demonstrate our approach on a large-scale dynamic vehicle routing problem with time windows.", "date": "2025-05-20", "categories": []} {"id": "2507.01037", "title": "Learning to Segment for Vehicle Routing Problems", "abstract": "Iterative heuristics are widely recognized as state-of-the-art for Vehicle Routing Problems (VRPs). In this work, we exploit a critical observation: a large portion of the solution remains stable, i.e., unchanged across search iterations, causing redundant computations, especially for large-scale VRPs with long subtours. To address this, we pioneer the formal study of the First-Segment-Then-Aggregate (FSTA) decomposition technique to accelerate iterative solvers. FSTA preserves stable solution segments during the search, aggregates nodes within each segment into fixed hypernodes, and focuses the search only on unstable portions. Yet, a key challenge lies in identifying which segments should be aggregated. To this end, we introduce Learning-to-Segment (L2Seg), a novel neural framework to intelligently differentiate potentially stable and unstable portions for FSTA decomposition. We present three L2Seg variants: non-autoregressive (globally comprehensive but locally indiscriminate), autoregressive (locally refined but globally deficient), and their synergy. Empirical results on CVRP and VRPTW show that L2Seg accelerates state-of-the-art solvers by 2x to 7x. We further provide in-depth analysis showing why synergy achieves the best performance. Notably, L2Seg is compatible with traditional, learning-based, and hybrid solvers, while supporting various VRPs.", "date": "2025-06-22", "categories": []} {"id": "2507.01076", "title": "Empirical Analysis Of Heuristic and Approximation Algorithms for the The Mutual-Visibility Problem", "abstract": "The NP-complete mutual-visibility (MV) problem currently lacks empirical analysis on its practical behaviour despite theoretical studies. This paper addresses this gap by implementing and evaluating three distinct algorithms -- a direct random heuristic, a hypergraph-based approximation, and a genetic algorithm -- on diverse synthetic graph datasets, including those with analytically known $\\mu(G)$ values and general graph models. Our results demonstrate that for smaller graphs, the algorithms consistently achieve MV set sizes aligning with theoretical bounds. However, for larger instances, achieved solution sizes notably diverge from theoretical limits; this, combined with the absence of tight bounds, complicates absolute quality assessment. Nevertheless, validation on known optimal graphs showed the Genetic Algorithm and other heuristics empirically performing best among tested methods.", "date": "2025-07-01", "categories": []} {"id": "2506.00533", "title": "RsGCN: Subgraph-Based Rescaling Enhances Generalization of GCNs for Solving Traveling Salesman Problems", "abstract": "GCN-based traveling salesman problem (TSP) solvers face two critical challenges: poor cross-scale generalization for TSPs and high training costs. To address these challenges, we propose a Subgraph-Based Rescaling Graph Convolutional Network (RsGCN). Focusing on the scale-dependent features (i.e., features varied with problem scales) related to nodes and edges, we design the subgraph-based rescaling to normalize edge lengths of subgraphs. Under a unified subgraph perspective, RsGCN can efficiently learn scale-generalizable representations from small-scale TSPs at low cost. To exploit and assess the heatmaps generated by RsGCN, we design a Reconstruction-Based Search (RBS), in which a reconstruction process based on adaptive weight is incorporated to help avoid local optima. Based on a combined architecture of RsGCN and RBS, our solver achieves remarkable generalization and low training cost: with only 3 epochs of training on a mixed-scale dataset containing instances with up to 100 nodes, it can be generalized successfully to 10K-node instances without any fine-tuning. Extensive experiments demonstrate our advanced performance across uniform-distribution instances of 9 different scales from 20 to 10K nodes and 78 real-world instances from TSPLIB, while requiring the fewest learnable parameters and training epochs among neural competitors.", "date": "2025-05-31", "categories": []} {"id": "2509.21902", "title": "DyRo-MCTS: A Robust Monte Carlo Tree Search Approach to Dynamic Job Shop Scheduling", "abstract": "Dynamic job shop scheduling, a fundamental combinatorial optimisation problem in various industrial sectors, poses substantial challenges for effective scheduling due to frequent disruptions caused by the arrival of new jobs. State-of-the-art methods employ machine learning to learn scheduling policies offline, enabling rapid responses to dynamic events. However, these offline policies are often imperfect, necessitating the use of planning techniques such as Monte Carlo Tree Search (MCTS) to improve performance at online decision time. The unpredictability of new job arrivals complicates online planning, as decisions based on incomplete problem information are vulnerable to disturbances. To address this issue, we propose the Dynamic Robust MCTS (DyRo-MCTS) approach, which integrates action robustness estimation into MCTS. DyRo-MCTS guides the production environment toward states that not only yield good scheduling outcomes but are also easily adaptable to future job arrivals. Extensive experiments show that DyRo-MCTS significantly improves the performance of offline-learned policies with negligible additional online planning time. Moreover, DyRo-MCTS consistently outperforms vanilla MCTS across various scheduling scenarios. Further analysis reveals that its ability to make robust scheduling decisions leads to long-term, sustainable performance gains under disturbances.", "date": "2025-09-26", "categories": []} {"id": "2509.18162", "title": "A Simple and Reproducible Hybrid Solver for a Truck-Drone VRP with Recharge", "abstract": "We study last-mile delivery with one truck and one drone under explicit battery management: the drone flies at twice the truck speed; each sortie must satisfy an endurance budget; after every delivery the drone recharges on the truck before the next launch. We introduce a hybrid reinforcement learning (RL) solver that couples an ALNS-based truck tour (with 2/3-opt and Or-opt) with a small pointer/attention policy that schedules drone sorties. The policy decodes launch-serve-rendezvous triplets with hard feasibility masks for endurance and post-delivery recharge; a fast, exact timeline simulator enforces launch/recovery handling and computes the true makespan used by masked greedy/beam decoding. On Euclidean instances with $N{=}50$, $E{=}0.7$, and $R{=}0.1$, the method achieves an average makespan of \\textbf{5.203}$\\pm$0.093, versus \\textbf{5.349}$\\pm$0.038 for ALNS and \\textbf{5.208}$\\pm$0.124 for NN -- i.e., \\textbf{2.73\\%} better than ALNS on average and within \\textbf{0.10\\%} of NN. Per-seed, the RL scheduler never underperforms ALNS on the same instance and ties or beats NN on two of three seeds. A decomposition of the makespan shows the expected truck-wait trade-off across heuristics; the learned scheduler balances both to minimize the total completion time. We provide a config-first implementation with plotting and significance-test utilities to support replication.", "date": "2025-09-17", "categories": []} {"id": "2507.04164", "title": "Structure As Search: Unsupervised Permutation Learning for Combinatorial Optimization", "abstract": "We propose a non-autoregressive framework for the Travelling Salesman Problem where solutions emerge directly from learned permutations, without requiring explicit search. By applying a similarity transformation to Hamiltonian cycles, the model learns to approximate permutation matrices via continuous relaxations. Our unsupervised approach achieves competitive performance against classical heuristics, demonstrating that the inherent structure of the problem can effectively guide combinatorial optimization without sequential decision-making. Our method offers concrete evidence that neural networks can directly capture and exploit combinatorial structure.", "date": "2025-07-05", "categories": []} {"id": "2509.16865", "title": "Large Language Models as End-to-end Combinatorial Optimization Solvers", "abstract": "Combinatorial optimization (CO) problems, central to decision-making scenarios like logistics and manufacturing, are traditionally solved using problem-specific algorithms requiring significant domain expertise. While large language models (LLMs) have shown promise in automating CO problem solving, existing approaches rely on intermediate steps such as code generation or solver invocation, limiting their generality and accessibility. This paper introduces a novel framework that empowers LLMs to serve as end-to-end CO solvers by directly mapping natural language problem descriptions to solutions. We propose a two-stage training strategy: supervised fine-tuning (SFT) imparts LLMs with solution generation patterns from domain-specific solvers, while a feasibility-and-optimality-aware reinforcement learning (FOARL) process explicitly mitigates constraint violations and refines solution quality. Evaluation across seven NP-hard CO problems shows that our method achieves a high feasibility rate and reduces the average optimality gap to 1.03-8.20% by tuning a 7B-parameter LLM, surpassing both general-purpose LLMs (e.g., GPT-4o), reasoning models (e.g., DeepSeek-R1), and domain-specific heuristics. Our method establishes a unified language-based pipeline for CO without extensive code execution or manual architectural adjustments for different problems, offering a general and language-driven alternative to traditional solver design while maintaining relative feasibility guarantees.", "date": "2025-09-21", "categories": []} {"id": "2509.16938", "title": "NeuFACO: Neural Focused Ant Colony Optimization for Traveling Salesman Problem", "abstract": "This study presents Neural Focused Ant Colony Optimization (NeuFACO), a non-autoregressive framework for the Traveling Salesman Problem (TSP) that combines advanced reinforcement learning with enhanced Ant Colony Optimization (ACO). NeuFACO employs Proximal Policy Optimization (PPO) with entropy regularization to train a graph neural network for instance-specific heuristic guidance, which is integrated into an optimized ACO framework featuring candidate lists, restricted tour refinement, and scalable local search. By leveraging amortized inference alongside ACO stochastic exploration, NeuFACO efficiently produces high-quality solutions across diverse TSP instances.", "date": "2025-09-21", "categories": []} {"id": "2509.19005", "title": "Quantum Annealing for Minimum Bisection Problem: A Machine Learning-based Approach for Penalty Parameter Tuning", "abstract": "The Minimum Bisection Problem is a well-known NP-hard problem in combinatorial optimization, with practical applications in areas such as parallel computing, network design, and machine learning. In this paper, we examine the potential of using D-Wave Systems' quantum annealing solvers to solve the Minimum Bisection Problem, which we formulate as a Quadratic Unconstrained Binary Optimization model. A key challenge in this formulation lies in choosing an appropriate penalty parameter, as it plays a crucial role in ensuring both the quality of the solution and the satisfaction of the problem's constraints. To address this, we introduce a novel machine learning-based approach for adaptive tuning of the penalty parameter. Specifically, we use a Gradient Boosting Regressor model trained to predict suitable penalty parameter values based on structural properties of the input graph, the number of nodes and the graph's density. This method enables the penalty parameter to be adjusted dynamically for each specific problem instance, improving the solver's ability to balance the competing goals of minimizing the cut size and maintaining equally sized partitions. We test our approach on a large dataset of randomly generated Erd\\H{o}s-R\\'enyi graphs with up to 4,000 nodes, and we compare the results with classical partitioning algorithms, Metis and Kernighan-Lin. Experimental findings demonstrate that our adaptive tuning strategy significantly improves the performance of the quantum annealing hybrid solver and consistently outperforms the classical methods used, indicating its potential as an alternative for the graph partitioning problem.", "date": "2025-09-23", "categories": []} {"id": "2504.06126", "title": "Accelerating Vehicle Routing via AI-Initialized Genetic Algorithms", "abstract": "Vehicle Routing Problems (VRP) are an extension of the Traveling Salesperson Problem and are a fundamental NP-hard challenge in combinatorial optimization. Solving VRP in real-time at large scale has become critical in numerous applications, from growing markets like last-mile delivery to emerging use-cases like interactive logistics planning. In many applications, one has to repeatedly solve VRP instances drawn from the same distribution, yet current state-of-the-art solvers treat each instance on its own without leveraging previous examples. We introduce an optimization framework where a reinforcement learning agent is trained on prior instances and quickly generates initial solutions, which are then further optimized by a genetic algorithm. This framework, Evolutionary Algorithm with Reinforcement Learning Initialization (EARLI), consistently outperforms current state-of-the-art solvers across various time budgets. For example, EARLI handles vehicle routing with 500 locations within one second, 10x faster than current solvers for the same solution quality, enabling real-time and interactive routing at scale. EARLI can generalize to new data, as we demonstrate on real e-commerce delivery data of a previously unseen city. By combining reinforcement learning and genetic algorithms, our hybrid framework takes a step forward to closer interdisciplinary collaboration between AI and optimization communities towards real-time optimization in diverse domains.", "date": "2025-04-08", "categories": []} {"id": "2509.15828", "title": "HyP-ASO: A Hybrid Policy-based Adaptive Search Optimization Framework for Large-Scale Integer Linear Programs", "abstract": "Directly solving large-scale Integer Linear Programs (ILPs) using traditional solvers is slow due to their NP-hard nature. While recent frameworks based on Large Neighborhood Search (LNS) can accelerate the solving process, their performance is often constrained by the difficulty in generating sufficiently effective neighborhoods. To address this challenge, we propose HyP-ASO, a hybrid policy-based adaptive search optimization framework that combines a customized formula with deep Reinforcement Learning (RL). The formula leverages feasible solutions to calculate the selection probabilities for each variable in the neighborhood generation process, and the RL policy network predicts the neighborhood size. Extensive experiments demonstrate that HyP-ASO significantly outperforms existing LNS-based approaches for large-scale ILPs. Additional experiments show it is lightweight and highly scalable, making it well-suited for solving large-scale ILPs.", "date": "2025-09-19", "categories": []} {"id": "2509.17870", "title": "Improving After-sales Service: Deep Reinforcement Learning for Dynamic Time Slot Assignment with Commitments and Customer Preferences", "abstract": "Problem definition: For original equipment manufacturers (OEMs), high-tech maintenance is a strategic component in after-sales services, involving close coordination between customers and service engineers. Each customer suggests several time slots for their maintenance task, from which the OEM must select one. This decision needs to be made promptly to support customers' planning. At the end of each day, routes for service engineers are planned to fulfill the tasks scheduled for the following day. We study this hierarchical and sequential decision-making problem-the Dynamic Time Slot Assignment Problem with Commitments and Customer Preferences (DTSAP-CCP)-in this paper. Methodology/results: Two distinct approaches are proposed: 1) an attention-based deep reinforcement learning with rollout execution (ADRL-RE) and 2) a scenario-based planning approach (SBP). The ADRL-RE combines a well-trained attention-based neural network with a rollout framework for online trajectory simulation. To support the training, we develop a neural heuristic solver that provides rapid route planning solutions, enabling efficient learning in complex combinatorial settings. The SBP approach samples several scenarios to guide the time slot assignment. Numerical experiments demonstrate the superiority of ADRL-RE and the stability of SBP compared to both rule-based and rollout-based approaches. Furthermore, the strong practicality of ADRL-RE is verified in a case study of after-sales service for large medical equipment. Implications: This study provides OEMs with practical decision-support tools for dynamic maintenance scheduling, balancing customer preferences and operational efficiency. In particular, our ADRL-RE shows strong real-world potential, supporting timely and customer-aligned maintenance scheduling.", "date": "2025-09-22", "categories": []} {"id": "2509.15275", "title": "Partial Column Generation with Graph Neural Networks for Team Formation and Routing", "abstract": "The team formation and routing problem is a challenging optimization problem with several real-world applications in fields such as airport, healthcare, and maintenance operations. To solve this problem, exact solution methods based on column generation have been proposed in the literature. In this paper, we propose a novel partial column generation strategy for settings with multiple pricing problems, based on predicting which ones are likely to yield columns with a negative reduced cost. We develop a machine learning model tailored to the team formation and routing problem that leverages graph neural networks for these predictions. Computational experiments demonstrate that applying our strategy enhances the solution method and outperforms traditional partial column generation approaches from the literature, particularly on hard instances solved under a tight time limit.", "date": "2025-09-18", "categories": []} {"id": "2509.15999", "title": "Inverse Optimization Latent Variable Models for Learning Costs Applied to Route Problems", "abstract": "Learning representations for solutions of constrained optimization problems (COPs) with unknown cost functions is challenging, as models like (Variational) Autoencoders struggle to enforce constraints when decoding structured outputs. We propose an Inverse Optimization Latent Variable Model (IO-LVM) that learns a latent space of COP cost functions from observed solutions and reconstructs feasible outputs by solving a COP with a solver in the loop. Our approach leverages estimated gradients of a Fenchel-Young loss through a non-differentiable deterministic solver to shape the latent space. Unlike standard Inverse Optimization or Inverse Reinforcement Learning methods, which typically recover a single or context-specific cost function, IO-LVM captures a distribution over cost functions, enabling the identification of diverse solution behaviors arising from different agents or conditions not available during the training process. We validate our method on real-world datasets of ship and taxi routes, as well as paths in synthetic graphs, demonstrating its ability to reconstruct paths and cycles, predict their distributions, and yield interpretable latent representations.", "date": "2025-09-19", "categories": []} {"id": "2509.14251", "title": "Unified Crew Planning and Replanning Optimization in Multi-Line Metro Systems Considering Workforce Heterogeneity", "abstract": "Metro crew planning is a key component of smart city development as it directly impacts the operational efficiency and service reliability of public transportation. With the rapid expansion of metro networks, effective multi-line scheduling and emergency management have become essential for large-scale seamless operations. However, current research focuses primarily on individual metro lines,with insufficient attention on cross-line coordination and rapid replanning during disruptions. Here, a unified optimization framework is presented for multi-line metro crew planning and replanning with heterogeneous workforce. Specifically, a hierarchical time-space network model is proposed to represent the unified crew action space, and computationally efficient constraints and formulations are derived for the crew's heterogeneous qualifications and preferences. Solution algorithms based on column generation and shortest path adjustment are further developed, utilizing the proposed network model. Experiments with real data from Shanghai and Beijing Metro demonstrate that the proposed methods outperform benchmark heuristics in both cost reduction and task completion,and achieve notable efficiency gains by incorporating cross-line operations, particularly for urgent tasks during disruptions. This work highlights the role of global optimization and cross-line coordination in multi-line metro system operations, providing insights into the efficient and reliable functioning of public transportation in smart cities.", "date": "2025-09-11", "categories": []} {"id": "2402.00041", "title": "Spatial-temporal-demand clustering for solving large-scale vehicle routing problems with time windows", "abstract": "Several metaheuristics use decomposition and pruning strategies to solve large-scale instances of the vehicle routing problem (VRP). Those complexity reduction techniques often rely on simple, problem-specific rules. However, the growth in available data and advances in computer hardware enable data-based approaches that use machine learning (ML) to improve scalability of solution algorithms. We propose a decompose-route-improve (DRI) framework that groups customers using clustering. Its similarity metric incorporates customers' spatial, temporal, and demand data and is formulated to reflect the problem's objective function and constraints. The resulting sub-routing problems can independently be solved using any suitable algorithm. We apply pruned local search (LS) between solved subproblems to improve the overall solution. Pruning is based on customers' similarity information obtained in the decomposition phase. In a computational study, we parameterize and compare existing clustering algorithms and benchmark the DRI against the Hybrid Genetic Search (HGS) of Vidal et al. (2013). Results show that our data-based approach outperforms classic cluster-first, route-second approaches solely based on customers' spatial information. The newly introduced similarity metric forms separate sub-VRPs and improves the selection of LS moves in the improvement phase. Thus, the DRI scales existing metaheuristics to achieve high-quality solutions faster for large-scale VRPs by efficiently reducing complexity. Further, the DRI can be easily adapted to various solution methods and VRP characteristics, such as distribution of customer locations and demands, depot location, and different time window scenarios, making it a generalizable approach to solving routing problems.", "date": "2024-01-20", "categories": []} {"id": "2502.12484", "title": "LocalEscaper: A Weakly-supervised Framework with Regional Reconstruction for Scalable Neural TSP Solvers", "abstract": "Neural solvers have shown significant potential in solving the Traveling Salesman Problem (TSP), yet current approaches face significant challenges. Supervised learning (SL)-based solvers require large amounts of high-quality labeled data, while reinforcement learning (RL)-based solvers, though less dependent on such data, often suffer from inefficiencies. To address these limitations, we propose LocalEscaper, a novel weakly-supervised learning framework for large-scale TSP. LocalEscaper effectively combines the advantages of both SL and RL, enabling effective training on datasets with low-quality labels. To further enhance solution quality, we introduce a regional reconstruction strategy, which is the key technique of this paper and mitigates the local-optima problem common in existing local reconstruction methods. Experimental results on both synthetic and real-world datasets demonstrate that LocalEscaper outperforms existing neural solvers, achieving remarkable results.", "date": "2025-02-18", "categories": []} {"id": "2308.14981", "title": "Sub-universal variational circuits for combinatorial optimization problems", "abstract": "Quantum variational circuits have gained significant attention due to their applications in the quantum approximate optimization algorithm and quantum machine learning research. This work introduces a novel class of classical probabilistic circuits designed for generating approximate solutions to combinatorial optimization problems constructed using two-bit stochastic matrices. Through a numerical study, we investigate the performance of our proposed variational circuits in solving the Max-Cut problem on various graphs of increasing sizes. Our classical algorithm demonstrates improved performance for several graph types to the quantum approximate optimization algorithm. Our findings suggest that evaluating the performance of quantum variational circuits against variational circuits with sub-universal gate sets is a valuable benchmark for identifying areas where quantum variational circuits can excel.", "date": "2023-08-29", "categories": []} {"id": "2509.13131", "title": "Reasoning with Preference Constraints: A Benchmark for Language Models in Many-to-One Matching Markets", "abstract": "Recent advances in reasoning with large language models (LLMs) have demonstrated strong performance on complex mathematical tasks, including combinatorial optimization. Techniques such as Chain-of-Thought and In-Context Learning have further enhanced this capability, making LLMs both powerful and accessible tools for a wide range of users, including non-experts. However, applying LLMs to matching problems, which require reasoning under preferential and structural constraints, remains underexplored. To address this gap, we introduce a novel benchmark of 369 instances of the College Admission Problem, a canonical example of a matching problem with preferences, to evaluate LLMs across key dimensions: feasibility, stability, and optimality. We employ this benchmark to assess the performance of several open-weight LLMs. Our results first reveal that while LLMs can satisfy certain constraints, they struggle to meet all evaluation criteria consistently. They also show that reasoning LLMs, like QwQ and GPT-oss, significantly outperform traditional models such as Llama, Qwen or Mistral, defined here as models used without any dedicated reasoning mechanisms. Moreover, we observed that LLMs reacted differently to the various prompting strategies tested, which include Chain-of-Thought, In-Context Learning and role-based prompting, with no prompt consistently offering the best performance. Finally, we report the performances from iterative prompting with auto-generated feedback and show that they are not monotonic; they can peak early and then significantly decline in later attempts. Overall, this work offers a new perspective on model reasoning performance and the effectiveness of prompting strategies in combinatorial optimization problems with preferential constraints.", "date": "2025-09-16", "categories": []} {"id": "2405.05978", "title": "Addressing Unboundedness in Quadratically-Constrained Mixed-Integer Problems", "abstract": "Mixed-integer (MI) quadratic models subject to quadratic constraints, known as All-Quadratic MI Programs, constitute a challenging class of NP-complete optimization problems. The particular scenario of unbounded integers defines a subclass that holds the distinction of being even undecidable [Jeroslow, 1973]. This complexity suggests a possible soft-spot for Mathematical Programming (MP) techniques, which otherwise constitute a good choice to treat MI problems. We consider the task of minimizing MI convex quadratic objective and constraint functions with unbounded decision variables. Given the theoretical weakness of white-box MP solvers to handle such models, we turn to black-box meta-heuristics of the Evolution Strategies (ESs) family, and question their capacity to solve this challenge. Through an empirical assessment of all-quadratic test-cases, across varying Hessian forms and condition numbers, we compare the performance of the CPLEX solver to modern MI ESs, which handle constraints by penalty. Our systematic investigation begins where the CPLEX solver encounters difficulties (timeouts as the search-space dimensionality increases, D>=30), and we report in detail on the D=64 case. Overall, the empirical observations confirm that black-box and white-box solvers can be competitive, where CPLEX is evidently outperformed on 13% of the cases. This trend is flipped when unboundedness is amplified by a significant translation of the optima, leading to a totally inferior performance of CPLEX at 83% of the cases. We also conclude that conditioning and separability are not intuitive factors in determining the hardness degree of this class of MI problems.", "date": "2024-05-06", "categories": []} {"id": "2406.15007", "title": "RouteFinder: Towards Foundation Models for Vehicle Routing Problems", "abstract": "This paper introduces RouteFinder, a comprehensive foundation model framework to tackle different Vehicle Routing Problem (VRP) variants. Our core idea is that a foundation model for VRPs should be able to represent variants by treating each as a subset of a generalized problem equipped with different attributes. We propose a unified VRP environment capable of efficiently handling any combination of these attributes. The RouteFinder model leverages a modern transformer-based encoder and global attribute embeddings to improve task representation. Additionally, we introduce two reinforcement learning techniques to enhance multi-task performance: mixed batch training, which enables training on different variants at once, and multi-variant reward normalization to balance different reward scales. Finally, we propose efficient adapter layers that enable fine-tuning for new variants with unseen attributes. Extensive experiments on 48 VRP variants show RouteFinder outperforms recent state-of-the-art learning methods. Our code is publicly available at https://github.com/ai4co/routefinder.", "date": "2024-06-21", "categories": []} {"id": "2509.09707", "title": "LLM-Based Instance-Driven Heuristic Bias In the Context of a Biased Random Key Genetic Algorithm", "abstract": "Integrating Large Language Models (LLMs) within metaheuristics opens a novel path for solving complex combinatorial optimization problems. While most existing approaches leverage LLMs for code generation to create or refine specific heuristics, they often overlook the structural properties of individual problem instances. In this work, we introduce a novel framework that integrates LLMs with a Biased Random-Key Genetic Algorithm (BRKGA) to solve the NP-hard Longest Run Subsequence problem. Our approach extends the instance-driven heuristic bias paradigm by introducing a human-LLM collaborative process to co-design and implement a set of computationally efficient metrics. The LLM analyzes these instance-specific metrics to generate a tailored heuristic bias, which steers the BRKGA toward promising areas of the search space. We conduct a comprehensive experimental evaluation, including rigorous statistical tests, convergence and behavioral analyses, and targeted ablation studies, comparing our method against a standard BRKGA baseline across 1,050 generated instances of varying complexity. Results show that our top-performing hybrid, BRKGA+Llama-4-Maverick, achieves statistically significant improvements over the baseline, particularly on the most complex instances. Our findings confirm that leveraging an LLM to produce an a priori, instance-driven heuristic bias is a valuable approach for enhancing metaheuristics in complex optimization domains.", "date": "2025-09-05", "categories": []} {"id": "2509.09933", "title": "Multi-Play Combinatorial Semi-Bandit Problem", "abstract": "In the combinatorial semi-bandit (CSB) problem, a player selects an action from a combinatorial action set and observes feedback from the base arms included in the action. While CSB is widely applicable to combinatorial optimization problems, its restriction to binary decision spaces excludes important cases involving non-negative integer flows or allocations, such as the optimal transport and knapsack problems.To overcome this limitation, we propose the multi-play combinatorial semi-bandit (MP-CSB), where a player can select a non-negative integer action and observe multiple feedbacks from a single arm in each round. We propose two algorithms for the MP-CSB. One is a Thompson-sampling-based algorithm that is computationally feasible even when the action space is exponentially large with respect to the number of arms, and attains $O(\\log T)$ distribution-dependent regret in the stochastic regime, where $T$ is the time horizon. The other is a best-of-both-worlds algorithm, which achieves $O(\\log T)$ variance-dependent regret in the stochastic regime and the worst-case $\\tilde{\\mathcal{O}}\\left( \\sqrt{T} \\right)$ regret in the adversarial regime. Moreover, its regret in adversarial one is data-dependent, adapting to the cumulative loss of the optimal action, the total quadratic variation, and the path-length of the loss sequence. Finally, we numerically show that the proposed algorithms outperform existing methods in the CSB literature.", "date": "2025-09-12", "categories": []} {"id": "2509.07019", "title": "An efficient deep reinforcement learning environment for flexible job-shop scheduling", "abstract": "The Flexible Job-shop Scheduling Problem (FJSP) is a classical combinatorial optimization problem that has a wide-range of applications in the real world. In order to generate fast and accurate scheduling solutions for FJSP, various deep reinforcement learning (DRL) scheduling methods have been developed. However, these methods are mainly focused on the design of DRL scheduling Agent, overlooking the modeling of DRL environment. This paper presents a simple chronological DRL environment for FJSP based on discrete event simulation and an end-to-end DRL scheduling model is proposed based on the proximal policy optimization (PPO). Furthermore, a short novel state representation of FJSP is proposed based on two state variables in the scheduling environment and a novel comprehensible reward function is designed based on the scheduling area of machines. Experimental results on public benchmark instances show that the performance of simple priority dispatching rules (PDR) is improved in our scheduling environment and our DRL scheduling model obtains competing performance compared with OR-Tools, meta-heuristic, DRL and PDR scheduling methods.", "date": "2025-09-07", "categories": []} {"id": "2504.08793", "title": "Constraint Programming Models For Serial Batch Scheduling With Minimum Batch Size", "abstract": "In serial batch (s-batch) scheduling, jobs are grouped in batches and processed sequentially within their batch. This paper considers multiple parallel machines, nonidentical job weights and release times, and sequence-dependent setup times between batches of different families. Although s-batch has been widely studied in the literature, very few papers have taken into account a minimum batch size, typical in practical settings such as semiconductor manufacturing and the metal industry. The problem with this minimum batch size requirement has been mostly tackled with dynamic programming and meta-heuristics, and no article has ever used constraint programming (CP) to do so. This paper fills this gap by proposing, three CP models for s-batching with minimum batch size: (i) an \\textit{Interval Assignment} model that computes and bounds the size of the batches using the presence literals of interval variables of the jobs. (ii) A \\textit{Global} model that exclusively uses global constraints that track the size of the batches over time. (iii) And a \\textit{Hybrid} model that combines the benefits of the extra global constraints with the efficiency of the sum-of-presences constraints to ensure the minimum batch sizes. The computational experiments on standard cases compare the three CP models with two existing mixed-integer programming (MIP) models from the literature. The results demonstrate the versatility of the proposed CP models to handle multiple variations of s-batching; and their ability to produce, in large instances, better solutions than the MIP models faster.", "date": "2025-04-07", "categories": []} {"id": "2508.02640", "title": "An Efficient Continuous-Time MILP for Integrated Aircraft Hangar Scheduling and Layout", "abstract": "Efficient management of aircraft MRO hangars requires the integration of spatial layout with time-continuous scheduling to minimize operational costs. We propose a continuous-time mixed-integer linear program that jointly optimizes aircraft placement and timing, overcoming the scalability limits of prior formulations. A comprehensive study benchmarks the model against a constructive heuristic, probes large-scale performance, and quantifies its sensitivity to temporal congestion. The model achieves orders-of-magnitude speedups on benchmarks from the literature, solving a long-standing congested instance in 0.11 seconds, and finds proven optimal solutions for instances with up to 40 aircraft. Within a one-hour limit for large-scale problems, the model finds solutions with small optimality gaps for instances up to 80 aircraft and provides strong bounds for problems with up to 160 aircraft. Optimized plans consistently increase hangar throughput (e.g., +33% serviced aircraft vs. a heuristic on instance RND-N030-I03), leading to lower delay penalties and higher asset utilization. These findings establish that exact optimization has become computationally viable for large-scale hangar planning, providing a validated tool that balances solution quality and computation time for strategic and operational decisions.", "date": "2025-08-04", "categories": []} {"id": "2509.05084", "title": "Recurrent State Encoders for Efficient Neural Combinatorial Optimization", "abstract": "The primary paradigm in Neural Combinatorial Optimization (NCO) are construction methods, where a neural network is trained to sequentially add one solution component at a time until a complete solution is constructed. We observe that the typical changes to the state between two steps are small, since usually only the node that gets added to the solution is removed from the state. An efficient model should be able to reuse computation done in prior steps. To that end, we propose to train a recurrent encoder that computes the state embeddings not only based on the state but also the embeddings of the step before. We show that the recurrent encoder can achieve equivalent or better performance than a non-recurrent encoder even if it consists of $3\\times$ fewer layers, thus significantly improving on latency. We demonstrate our findings on three different problems: the Traveling Salesman Problem (TSP), the Capacitated Vehicle Routing Problem (CVRP), and the Orienteering Problem (OP) and integrate the models into a large neighborhood search algorithm, to showcase the practical relevance of our findings.", "date": "2025-09-05", "categories": []} {"id": "2402.04915", "title": "Moco: A Learnable Meta Optimizer for Combinatorial Optimization", "abstract": "Relevant combinatorial optimization problems (COPs) are often NP-hard. While they have been tackled mainly via handcrafted heuristics in the past, advances in neural networks have motivated the development of general methods to learn heuristics from data. Many approaches utilize a neural network to directly construct a solution, but are limited in further improving based on already constructed solutions at inference time. Our approach, Moco, defines a lightweight solution construction procedure, guided by a single continuous vector $\\theta$ (called heatmap) and learns a neural network to update $\\theta$ for a single instance of a COP at inference time. The update is based on various features of the current search state. The training procedure is budget aware, targeting the overall best solution found during the entire search. Moco is a fully learnable meta optimizer not utilizing problem specific heuristics or requiring optimal solutions for training. We test Moco on the Traveling Salesman Problem (TSP) and Maximum Independent Set (MIS) and show that it significantly improves over other heatmap based methods.", "date": "2024-02-07", "categories": []} {"id": "2504.04421", "title": "Deliberate Planning of 3D Bin Packing on Packing Configuration Trees", "abstract": "Online 3D Bin Packing Problem (3D-BPP) has widespread applications in industrial automation. Existing methods usually solve the problem with limited resolution of spatial discretization, and/or cannot deal with complex practical constraints well. We propose to enhance the practical applicability of online 3D-BPP via learning on a novel hierarchical representation, packing configuration tree (PCT). PCT is a full-fledged description of the state and action space of bin packing which can support packing policy learning based on deep reinforcement learning (DRL). The size of the packing action space is proportional to the number of leaf nodes, making the DRL model easy to train and well-performing even with continuous solution space. We further discover the potential of PCT as tree-based planners in deliberately solving packing problems of industrial significance, including large-scale packing and different variations of BPP setting. A recursive packing method is proposed to decompose large-scale packing into smaller sub-trees while a spatial ensemble mechanism integrates local solutions into global. For different BPP variations with additional decision variables, such as lookahead, buffering, and offline packing, we propose a unified planning framework enabling out-of-the-box problem solving. Extensive evaluations demonstrate that our method outperforms existing online BPP baselines and is versatile in incorporating various practical constraints. The planning process excels across large-scale problems and diverse problem variations. We develop a real-world packing robot for industrial warehousing, with careful designs accounting for constrained placement and transportation stability. Our packing robot operates reliably and efficiently on unprotected pallets at 10 seconds per box. It achieves averagely 19 boxes per pallet with 57.4% space utilization for relatively large-size boxes.", "date": "2025-04-06", "categories": []} {"id": "2509.04007", "title": "AutoPBO: LLM-powered Optimization for Local Search PBO Solvers", "abstract": "Pseudo-Boolean Optimization (PBO) provides a powerful framework for modeling combinatorial problems through pseudo-Boolean (PB) constraints. Local search solvers have shown excellent performance in PBO solving, and their efficiency is highly dependent on their internal heuristics to guide the search. Still, their design often requires significant expert effort and manual tuning in practice. While Large Language Models (LLMs) have demonstrated potential in automating algorithm design, their application to optimizing PBO solvers remains unexplored. In this work, we introduce AutoPBO, a novel LLM-powered framework to automatically enhance PBO local search solvers. We conduct experiments on a broad range of four public benchmarks, including one real-world benchmark, a benchmark from PB competition, an integer linear programming optimization benchmark, and a crafted combinatorial benchmark, to evaluate the performance improvement achieved by AutoPBO and compare it with six state-of-the-art competitors, including two local search PBO solvers NuPBO and OraSLS, two complete PB solvers PBO-IHS and RoundingSat, and two mixed integer programming (MIP) solvers Gurobi and SCIP. AutoPBO demonstrates significant improvements over previous local search approaches, while maintaining competitive performance compared to state-of-the-art competitors. The results suggest that AutoPBO offers a promising approach to automating local search solver design.", "date": "2025-09-04", "categories": []} {"id": "2509.04089", "title": "Gromov-Wasserstein and optimal transport: from assignment problems to probabilistic numeric", "abstract": "The assignment problem, a cornerstone of operations research, seeks an optimal one-to-one mapping between agents and tasks to minimize total cost. This work traces its evolution from classical formulations and algorithms to modern optimal transport (OT) theory, positioning the Quadratic Assignment Problem (QAP) and related structural matching tasks within this framework. We connect the linear assignment problem to Monge's transport problem, Kantorovich's relaxation, and Wasserstein distances, then extend to cases where source and target lie in different metric-measure spaces requiring Gromov-Wasserstein (GW) distances. GW formulations, including the fused GW variant that integrates structural and feature information, naturally address QAP-like problems by optimizing alignment based on both intra-domain distances and cross-domain attributes. Applications include graph matching, keypoint correspondence, and feature-based assignments. We present exact solvers, Genetic Algorithms (GA), and multiple GW variants, including a proposed multi-initialization strategy (GW-MultiInit) that mitigates the risk of getting stuck in local optima alongside entropic Sinkhorn-based approximations and fused GW. Computational experiments on capacitated QAP instances show that GW-MultiInit consistently achieves near-optimal solutions and scales efficiently to large problems where exact methods become impractical, while parameterized EGW and FGW variants provide flexible trade-offs between accuracy and runtime. Our findings provide theoretical foundations, computational insights, and practical guidelines for applying OT and GW methods to QAP and other real-world matching problems, such as those in machine learning and logistics.", "date": "2025-09-04", "categories": []} {"id": "2507.09879", "title": "Covering a Few Submodular Constraints and Applications", "abstract": "We consider the problem of covering multiple submodular constraints. Given a finite ground set $N$, a cost function $c: N \\rightarrow \\mathbb{R}_+$, $r$ monotone submodular functions $f_1,f_2,\\ldots,f_r$ over $N$ and requirements $b_1,b_2,\\ldots,b_r$ the goal is to find a minimum cost subset $S \\subseteq N$ such that $f_i(S) \\ge b_i$ for $1 \\le i \\le r$. When $r=1$ this is the well-known Submodular Set Cover problem. Previous work \\cite{chekuri2022covering} considered the setting when $r$ is large and developed bi-criteria approximation algorithms, and approximation algorithms for the important special case when each $f_i$ is a weighted coverage function. These are fairly general models and capture several concrete and interesting problems as special cases. The approximation ratios for these problem are at least $\\Omega(\\log r)$ which is unavoidable when $r$ is part of the input. In this paper, motivated by some recent applications, we consider the problem when $r$ is a \\emph{fixed constant} and obtain two main results. For covering multiple submodular constraints we obtain a randomized bi-criteria approximation algorithm that for any given integer $\\alpha \\ge 1$ outputs a set $S$ such that $f_i(S) \\ge$ $(1-1/e^\\alpha -\\epsilon)b_i$ for each $i \\in [r]$ and $\\mathbb{E}[c(S)] \\le (1+\\epsilon)\\alpha \\cdot \\sf{OPT}$. Second, when the $f_i$ are weighted coverage functions from a deletion-closed set system we obtain a $(1+\\epsilon)$ $(\\frac{e}{e-1})$ $(1+\\beta)$-approximation where $\\beta$ is the approximation ratio for the underlying set cover instances via the natural LP. These results show that one can obtain nearly as good an approximation for any fixed $r$ as what one would achieve for $r=1$. We mention some applications that follow easily from these general results and anticipate more in the future.", "date": "2025-07-14", "categories": []} {"id": "2509.02782", "title": "Key Principles in Cross-Domain Hyper-Heuristic Performance", "abstract": "Cross-domain selection hyper-heuristics aim to distill decades of research on problem-specific heuristic search algorithms into adaptable general-purpose search strategies. In this respect, existing selection hyper-heuristics primarily focus on an adaptive selection of low-level heuristics (LLHs) from a predefined set. In contrast, we concentrate on the composition of this set and its strategic transformations. We systematically analyze transformations based on three key principles: solution acceptance, LLH repetitions, and perturbation intensity, i.e., the proportion of a solution affected by a perturbative LLH. We demonstrate the raw effects of our transformations on a trivial unbiased random selection mechanism. With an appropriately constructed transformation, this trivial method outperforms all available state-of-the-art hyper-heuristics on three challenging real-world domains and finds 11 new best-known solutions. The same method is competitive with the winner of the CHeSC competition, commonly used as the standard cross-domain benchmark. Moreover, we accompany several recent hyper-heuristics with such strategic transformations. Using this approach, we outperform the current state-of-the-art methods on both the CHeSC benchmark and real-world domains while often simplifying their designs.", "date": "2025-09-02", "categories": []} {"id": "2509.00099", "title": "LLM-QUBO: An End-to-End Framework for Automated QUBO Transformation from Natural Language Problem Descriptions", "abstract": "Quantum annealing offers a promising paradigm for solving NP-hard combinatorial optimization problems, but its practical application is severely hindered by two challenges: the complex, manual process of translating problem descriptions into the requisite Quadratic Unconstrained Binary Optimization (QUBO) format and the scalability limitations of current quantum hardware. To address these obstacles, we propose a novel end-to-end framework, LLM-QUBO, that automates this entire formulation-to-solution pipeline. Our system leverages a Large Language Model (LLM) to parse natural language, automatically generating a structured mathematical representation. To overcome hardware limitations, we integrate a hybrid quantum-classical Benders' decomposition method. This approach partitions the problem, compiling the combinatorial complex master problem into a compact QUBO format, while delegating linearly structured sub-problems to classical solvers. The correctness of the generated QUBO and the scalability of the hybrid approach are validated using classical solvers, establishing a robust performance baseline and demonstrating the framework's readiness for quantum hardware. Our primary contribution is a synergistic computing paradigm that bridges classical AI and quantum computing, addressing key challenges in the practical application of optimization problem. This automated workflow significantly reduces the barrier to entry, providing a viable pathway to transform quantum devices into accessible accelerators for large-scale, real-world optimization challenges.", "date": "2025-08-27", "categories": []} {"id": "2502.14536", "title": "Algorithms for the preordering problem and their application to the task of jointly clustering and ordering the accounts of a social network", "abstract": "The NP-hard maximum value preordering problem is both a joint relaxation and a hybrid of the clique partition problem (a clustering problem) and the partial ordering problem. Toward approximate solutions and lower bounds, we introduce a linear-time 4-approximation algorithm that constructs a maximum dicut of a subgraph and define local search heuristics. Toward upper bounds, we tighten a linear program relaxation by the class of odd closed walk inequalities that define facets, as we show, of the preorder polytope. We contribute implementations of the algorithms, apply these to the task of jointly clustering and partially ordering the accounts of published social networks, and compare the output and efficiency qualitatively and quantitatively.", "date": "2025-02-20", "categories": []} {"id": "2508.20953", "title": "A Multi-Objective Genetic Algorithm for Healthcare Workforce Scheduling", "abstract": "Workforce scheduling in the healthcare sector is a significant operational challenge, characterized by fluctuating patient loads, diverse clinical skills, and the critical need to control labor costs while upholding high standards of patient care. This problem is inherently multi-objective, demanding a delicate balance between competing goals: minimizing payroll, ensuring adequate staffing for patient needs, and accommodating staff preferences to mitigate burnout. We propose a Multi-objective Genetic Algorithm (MOO-GA) that models the hospital unit workforce scheduling problem as a multi-objective optimization task. Our model incorporates real-world complexities, including hourly appointment-driven demand and the use of modular shifts for a multi-skilled workforce. By defining objective functions for cost, patient care coverage, and staff satisfaction, the GA navigates the vast search space to identify a set of high-quality, non-dominated solutions. Demonstrated on datasets representing a typical hospital unit, the results show that our MOO-GA generates robust and balanced schedules. On average, the schedules produced by our algorithm showed a 66\\% performance improvement over a baseline that simulates a conventional, manual scheduling process. This approach effectively manages trade-offs between critical operational and staff-centric objectives, providing a practical decision support tool for nurse managers and hospital administrators.", "date": "2025-08-28", "categories": []} {"id": "2503.11435", "title": "Preference Elicitation for Multi-objective Combinatorial Optimization with Active Learning and Maximum Likelihood Estimation", "abstract": "Real-life combinatorial optimization problems often involve several conflicting objectives, such as price, product quality and sustainability. A computationally-efficient way to tackle multiple objectives is to aggregate them into a single-objective function, such as a linear combination. However, defining the weights of the linear combination upfront is hard; alternatively, the use of interactive learning methods that ask users to compare candidate solutions is highly promising. The key challenges are to generate candidates quickly, to learn an objective function that leads to high-quality solutions and to do so with few user interactions. We build upon the Constructive Preference Elicitation framework and show how each of the three properties can be improved: to increase the interaction speed we investigate using pools of (relaxed) solutions, to improve the learning we adopt Maximum Likelihood Estimation of a Bradley-Terry preference model; and to reduce the number of user interactions, we select the pair of candidates to compare with an ensemble-based acquisition function inspired from Active Learning. Our careful experimentation demonstrates each of these improvements: on a PC configuration task and a realistic multi-instance routing problem, our method selects queries faster, needs fewer queries and synthesizes higher-quality combinatorial solutions than previous CPE methods.", "date": "2025-03-14", "categories": []} {"id": "2508.19737", "title": "InfraredGP: Efficient Graph Partitioning via Spectral Graph Neural Networks with Negative Corrections", "abstract": "Graph partitioning (GP), a.k.a. community detection, is a classic problem that divides nodes of a graph into densely-connected blocks. From a perspective of graph signal processing, we find that graph Laplacian with a negative correction can derive graph frequencies beyond the conventional range $[0, 2]$. To explore whether the low-frequency information beyond this range can encode more informative properties about community structures, we propose InfraredGP. It (\\romannumeral1) adopts a spectral GNN as its backbone combined with low-pass filters and a negative correction mechanism, (\\romannumeral2) only feeds random inputs to this backbone, (\\romannumeral3) derives graph embeddings via one feed-forward propagation (FFP) without any training, and (\\romannumeral4) obtains feasible GP results by feeding the derived embeddings to BIRCH. Surprisingly, our experiments demonstrate that based solely on the negative correction mechanism that amplifies low-frequency information beyond $[0, 2]$, InfraredGP can derive distinguishable embeddings for some standard clustering modules (e.g., BIRCH) and obtain high-quality results for GP without any training. Following the IEEE HPEC Graph Challenge benchmark, we evaluate InfraredGP for both static and streaming GP, where InfraredGP can achieve much better efficiency (e.g., 16x-23x faster) and competitive quality over various baselines. We have made our code public at https://github.com/KuroginQin/InfraredGP", "date": "2025-08-27", "categories": []} {"id": "2508.19963", "title": "Flocking Behavior: An Innovative Inspiration for the Optimization of Production Plants", "abstract": "Optimizing modern production plants using the job-shop principle is a known hard problem. For very large plants, like semiconductor fabs, the problem becomes unsolvable on a plant-wide scale in a reasonable amount of time using classical linear optimization. An alternative approach is the use of swarm intelligence algorithms. These have been applied to the job-shop problem before, but often in a centrally calculated way where they are applied to the solution space, but they can be implemented in a bottom-up fashion to avoid global result computation as well. One of the problems in semiconductor production is that the production process requires a lot of switching between machines that process lots one after the other and machines that process batches of lots at once, often with long processing times. In this paper, we address this switching problem with the ``boids'' flocking algorithm that was originally used in robotics and movie industry. The flocking behavior is a bio-inspired algorithm that uses only local information and interaction based on simple heuristics. We show that this algorithm addresses these valid considerations in production plant optimization, as it reacts to the switching of machine kinds similar to how a swarm of flocking animals would react to obstacles in its course.", "date": "2025-08-27", "categories": []} {"id": "2508.19218", "title": "The Subset Sum Matching Problem", "abstract": "This paper presents a new combinatorial optimisation task, the Subset Sum Matching Problem (SSMP), which is an abstraction of common financial applications such as trades reconciliation. We present three algorithms, two suboptimal and one optimal, to solve this problem. We also generate a benchmark to cover different instances of SSMP varying in complexity, and carry out an experimental evaluation to assess the performance of the approaches.", "date": "2025-08-26", "categories": []} {"id": "2409.10575", "title": "A Tie-breaking based Local Search Algorithm for Stable Matching Problems", "abstract": "The stable marriage problem with incomplete lists and ties (SMTI) and the hospitals/residents problem with ties (HRT) are important in matching theory with broad practical applications. In this paper, we introduce a tie-breaking based local search (TBLS) algorithm designed to achieve a weakly stable matching of maximum size for both the SMTI and HRT problems. TBLS begins by arbitrarily resolving all ties and iteratively refines the tie-breaking strategy by adjusting the relative order within ties based on preference ranks and the current stable matching. Additionally, we introduce TBLS-E, an equity-focused variant of TBLS, specifically designed for the SMTI problem. This variant maintains the objective of maximizing matching size, while enhancing equity through two simple modifications. In comparison with ten other approximation and local search algorithms, TBLS achieves the highest matching size, while TBLS-E exhibits the lowest sex equality cost. Significantly, TBLS-E preserves a matching size comparable to that of TBLS. Both our algorithms demonstrate faster computational speed than other local search algorithms in solving large-scale instances. Moreover, our scalability analysis shows that both algorithms maintain efficient performance as problem size increases.", "date": "2024-09-15", "categories": []} {"id": "2502.06777", "title": "Learning an Optimal Assortment Policy under Observational Data", "abstract": "We study the fundamental problem of offline assortment optimization under the Multinomial Logit (MNL) model, where sellers must determine the optimal subset of the products to offer based solely on historical customer choice data. While most existing approaches to learning-based assortment optimization focus on the online learning of the optimal assortment through repeated interactions with customers, such exploration can be costly or even impractical in many real-world settings. In this paper, we consider the offline learning paradigm and investigate the minimal data requirements for efficient offline assortment optimization. To this end, we introduce Pessimistic Rank-Breaking (PRB), an algorithm that combines rank-breaking with pessimistic estimation. We prove that PRB is nearly minimax optimal by establishing the tight suboptimality upper bound and a nearly matching lower bound. This further shows that \"optimal item coverage\" - where each item in the optimal assortment appears sufficiently often in the historical data - is both sufficient and necessary for efficient offline learning. This significantly relaxes the previous requirement of observing the complete optimal assortment in the data. Our results provide fundamental insights into the data requirements for offline assortment optimization under the MNL model.", "date": "2025-02-10", "categories": []} {"id": "2505.05119", "title": "USPR: Learning a Unified Solver for Profiled Routing", "abstract": "The Profiled Vehicle Routing Problem (PVRP) extends the classical VRP by incorporating vehicle-client-specific preferences and constraints, reflecting real-world requirements such as zone restrictions and service-level preferences. While recent reinforcement-learning solvers have shown promising performance, they require retraining for each new profile distribution, suffer from poor representation ability, and struggle to generalize to out-of-distribution instances. In this paper, we address these limitations by introducing Unified Solver for Profiled Routing (USPR), a novel framework that natively handles arbitrary profile types. USPR introduces on three key innovations: (i) Profile Embeddings (PE) to encode any combination of profile types; (ii) Multi-Head Profiled Attention (MHPA), an attention mechanism that models rich interactions between vehicles and clients; (iii) Profile-aware Score Reshaping (PSR), which dynamically adjusts decoder logits using profile scores to improve generalization. Empirical results on diverse PVRP benchmarks demonstrate that USPR achieves state-of-the-art results among learning-based methods while offering significant gains in flexibility and computational efficiency. We make our source code publicly available to foster future research.", "date": "2025-05-08", "categories": []} {"id": "2508.06847", "title": "MOCA-HESP: Meta High-dimensional Bayesian Optimization for Combinatorial and Mixed Spaces via Hyper-ellipsoid Partitioning", "abstract": "High-dimensional Bayesian Optimization (BO) has attracted significant attention in recent research. However, existing methods have mainly focused on optimizing in continuous domains, while combinatorial (ordinal and categorical) and mixed domains still remain challenging. In this paper, we first propose MOCA-HESP, a novel high-dimensional BO method for combinatorial and mixed variables. The key idea is to leverage the hyper-ellipsoid space partitioning (HESP) technique with different categorical encoders to work with high-dimensional, combinatorial and mixed spaces, while adaptively selecting the optimal encoders for HESP using a multi-armed bandit technique. Our method, MOCA-HESP, is designed as a \\textit{meta-algorithm} such that it can incorporate other combinatorial and mixed BO optimizers to further enhance the optimizers' performance. Finally, we develop three practical BO methods by integrating MOCA-HESP with state-of-the-art BO optimizers for combinatorial and mixed variables: standard BO, CASMOPOLITAN, and Bounce. Our experimental results on various synthetic and real-world benchmarks show that our methods outperform existing baselines. Our code implementation can be found at https://github.com/LamNgo1/moca-hesp", "date": "2025-08-09", "categories": []} {"id": "2508.13020", "title": "e-boost: Boosted E-Graph Extraction with Adaptive Heuristics and Exact Solving", "abstract": "E-graphs have attracted growing interest in many fields, particularly in logic synthesis and formal verification. E-graph extraction is a challenging NP-hard combinatorial optimization problem. It requires identifying optimal terms from exponentially many equivalent expressions, serving as the primary performance bottleneck in e-graph based optimization tasks. However, traditional extraction methods face a critical trade-off: heuristic approaches offer speed but sacrifice optimality, while exact methods provide optimal solutions but face prohibitive computational costs on practical problems. We present e-boost, a novel framework that bridges this gap through three key innovations: (1) parallelized heuristic extraction that leverages weak data dependence to compute DAG costs concurrently, enabling efficient multi-threaded performance without sacrificing extraction quality; (2) adaptive search space pruning that employs a parameterized threshold mechanism to retain only promising candidates, dramatically reducing the solution space while preserving near-optimal solutions; and (3) initialized exact solving that formulates the reduced problem as an Integer Linear Program with warm-start capabilities, guiding solvers toward high-quality solutions faster.\n Across the diverse benchmarks in formal verification and logic synthesis fields, e-boost demonstrates 558x runtime speedup over traditional exact approaches (ILP) and 19.04% performance improvement over the state-of-the-art extraction framework (SmoothE). In realistic logic synthesis tasks, e-boost produces 7.6% and 8.1% area improvements compared to conventional synthesis tools with two different technology mapping libraries. e-boost is available at https://github.com/Yu-Maryland/e-boost.", "date": "2025-08-18", "categories": []} {"id": "2508.17087", "title": "Solving the Min-Max Multiple Traveling Salesmen Problem via Learning-Based Path Generation and Optimal Splitting", "abstract": "This study addresses the Min-Max Multiple Traveling Salesmen Problem ($m^3$-TSP), which aims to coordinate tours for multiple salesmen such that the length of the longest tour is minimized. Due to its NP-hard nature, exact solvers become impractical under the assumption that $P \\ne NP$. As a result, learning-based approaches have gained traction for their ability to rapidly generate high-quality approximate solutions. Among these, two-stage methods combine learning-based components with classical solvers, simplifying the learning objective. However, this decoupling often disrupts consistent optimization, potentially degrading solution quality. To address this issue, we propose a novel two-stage framework named \\textbf{Generate-and-Split} (GaS), which integrates reinforcement learning (RL) with an optimal splitting algorithm in a joint training process. The splitting algorithm offers near-linear scalability with respect to the number of cities and guarantees optimal splitting in Euclidean space for any given path. To facilitate the joint optimization of the RL component with the algorithm, we adopt an LSTM-enhanced model architecture to address partial observability. Extensive experiments show that the proposed GaS framework significantly outperforms existing learning-based approaches in both solution quality and transferability.", "date": "2025-08-23", "categories": []} {"id": "2508.17452", "title": "ReviBranch: Deep Reinforcement Learning for Branch-and-Bound with Revived Trajectories", "abstract": "The Branch-and-bound (B&B) algorithm is the main solver for Mixed Integer Linear Programs (MILPs), where the selection of branching variable is essential to computational efficiency. However, traditional heuristics for branching often fail to generalize across heterogeneous problem instances, while existing learning-based methods such as imitation learning (IL) suffers from dependence on expert demonstration quality, and reinforcement learning (RL) struggles with limitations in sparse rewards and dynamic state representation challenges. To address these issues, we propose ReviBranch, a novel deep RL framework that constructs revived trajectories by reviving explicit historical correspondences between branching decisions and their corresponding graph states along search-tree paths. During training, ReviBranch enables agents to learn from complete structural evolution and temporal dependencies within the branching process. Additionally, we introduce an importance-weighted reward redistribution mechanism that transforms sparse terminal rewards into dense stepwise feedback, addressing the sparse reward challenge. Extensive experiments on different MILP benchmarks demonstrate that ReviBranch outperforms state-of-the-art RL methods, reducing B&B nodes by 4.0% and LP iterations by 2.2% on large-scale instances. The results highlight the robustness and generalizability of ReviBranch across heterogeneous MILP problem classes.", "date": "2025-08-24", "categories": []} {"id": "2406.01661", "title": "A Diffusion Model Framework for Unsupervised Neural Combinatorial Optimization", "abstract": "Learning to sample from intractable distributions over discrete sets without relying on corresponding training data is a central problem in a wide range of fields, including Combinatorial Optimization. Currently, popular deep learning-based approaches rely primarily on generative models that yield exact sample likelihoods. This work introduces a method that lifts this restriction and opens the possibility to employ highly expressive latent variable models like diffusion models. Our approach is conceptually based on a loss that upper bounds the reverse Kullback-Leibler divergence and evades the requirement of exact sample likelihoods. We experimentally validate our approach in data-free Combinatorial Optimization and demonstrate that our method achieves a new state-of-the-art on a wide range of benchmark problems.", "date": "2024-06-03", "categories": []} {"id": "2504.04310", "title": "CO-Bench: Benchmarking Language Model Agents in Algorithm Search for Combinatorial Optimization", "abstract": "Although LLM-based agents have attracted significant attention in domains such as software engineering and machine learning research, their role in advancing combinatorial optimization (CO) remains relatively underexplored. This gap underscores the need for a deeper understanding of their potential in tackling structured, constraint-intensive problems -- a pursuit currently limited by the absence of comprehensive benchmarks for systematic investigation. To address this, we introduce CO-Bench, a benchmark suite featuring 36 real-world CO problems drawn from a broad range of domains and complexity levels. CO-Bench includes structured problem formulations and curated data to support rigorous investigation of LLM agents. We evaluate multiple agentic frameworks against established human-designed algorithms, revealing the strengths and limitations of existing LLM agents and identifying promising directions for future research. CO-Bench is publicly available at https://github.com/sunnweiwei/CO-Bench.", "date": "2025-04-06", "categories": []} {"id": "2508.16171", "title": "SPL-LNS: Sampling-Enhanced Large Neighborhood Search for Solving Integer Linear Programs", "abstract": "Large Neighborhood Search (LNS) is a common heuristic in combinatorial optimization that iteratively searches over a large neighborhood of the current solution for a better one. Recently, neural network-based LNS solvers have achieved great success in solving Integer Linear Programs (ILPs) by learning to greedily predict the locally optimal solution for the next neighborhood proposal. However, this greedy approach raises two key concerns: (1) to what extent this greedy proposal suffers from local optima, and (2) how can we effectively improve its sample efficiency in the long run. To address these questions, this paper first formulates LNS as a stochastic process, and then introduces SPL-LNS, a sampling-enhanced neural LNS solver that leverages locally-informed proposals to escape local optima. We also develop a novel hindsight relabeling method to efficiently train SPL-LNS on self-generated data. Experimental results demonstrate that SPL-LNS substantially surpasses prior neural LNS solvers for various ILP problems of different sizes.", "date": "2025-08-22", "categories": []} {"id": "2503.20001", "title": "Unsupervised Learning for Quadratic Assignment", "abstract": "We introduce PLUME search, a data-driven framework that enhances search efficiency in combinatorial optimization through unsupervised learning. Unlike supervised or reinforcement learning, PLUME search learns directly from problem instances using a permutation-based loss with a non-autoregressive approach. We evaluate its performance on the quadratic assignment problem, a fundamental NP-hard problem that encompasses various combinatorial optimization problems. Experimental results demonstrate that PLUME search consistently improves solution quality. Furthermore, we study the generalization behavior and show that the learned model generalizes across different densities and sizes.", "date": "2025-03-25", "categories": []} {"id": "2508.14071", "title": "Edge-Selector Model Applied for Local Search Neighborhood for Solving Vehicle Routing Problems", "abstract": "This research proposes a hybrid Machine Learning and metaheuristic mechanism that is designed to solve Vehicle Routing Problems (VRPs). The main of our method is an edge solution selector model, which classifies solution edges to identify prohibited moves during the local search, hence guiding the search process within metaheuristic baselines. Two learning-based mechanisms are used to develop the edge selector: a simple tabular binary classifier and a Graph Neural Network (GNN). The tabular classifier employs Gradient Boosting Trees and Feedforward Neural Network as the baseline algorithms. Adjustments to the decision threshold are also applied to handle the class imbalance in the problem instance. An alternative mechanism employs the GNN to utilize graph structure for direct solution edge prediction, with the objective of guiding local search by predicting prohibited moves. These hybrid mechanisms are then applied in state-fo-the-art metaheuristic baselines. Our method demonstrates both scalability and generalizability, achieving performance improvements across different baseline metaheuristics, various problem sizes and variants, including the Capacitated Vehicle Routing Problem (CVRP) and CVRP with Time Windows (CVRPTW). Experimental evaluations on benchmark datasets up to 30,000 customer nodes, supported by pair-wise statistical analysis, verify the observed improvements.", "date": "2025-08-12", "categories": []} {"id": "2504.02383", "title": "Reinforcement Learning for Solving the Pricing Problem in Column Generation: Applications to Vehicle Routing", "abstract": "In this paper, we address the problem of Column Generation (CG) using Reinforcement Learning (RL). Specifically, we use a RL model based on the attention-mechanism architecture to find the columns with most negative reduced cost in the Pricing Problem (PP). Unlike previous Machine Learning (ML) applications for CG, our model deploys an end-to-end mechanism as it independently solves the pricing problem without the help of any heuristic. We consider a variant of Vehicle Routing Problem (VRP) as a case study for our method. Through a set of experiments where our method is compared against a Dynamic Programming (DP)-based heuristic for solving the PP, we show that our method solves the linear relaxation up to a reasonable objective gap in significantly shorter running times.", "date": "2025-04-03", "categories": []} {"id": "2508.13437", "title": "Discrete Optimization of Min-Max Violation and its Applications Across Computational Sciences", "abstract": "We introduce the Discrete Min-Max Violation (DMMV) as a general optimization problem which seeks an assignment of discrete values to variables that minimizes the largest constraint violation. This context-free mathematical formulation is applicable to a wide range of use cases that have worst-case performance requirements. After defining the DMMV problem mathematically, we explore its properties to establish a foundational understanding. To tackle DMMV instance sizes of practical relevance, we develop a GPU-accelerated heuristic that takes advantage of the mathematical properties of DMMV for speeding up the solution process. We demonstrate the versatile applicability of our heuristic by solving three optimization problems as use cases: (1) post-training quantization of language models, (2) discrete tomography, and (3) Finite Impulse Response (FIR) filter design. In quantization without outlier separation, our heuristic achieves 14% improvement on average over existing methods. In discrete tomography, it reduces reconstruction error by 16% under uniform noise and accelerates computations by a factor of 6 on GPU. For FIR filter design, it nearly achieves 50% ripple reduction compared to using the commercial integer optimization solver, Gurobi. Our comparative results point to the benefits of studying DMMV as a context-free optimization problem and the advantages that our proposed heuristic offers on three distinct problems. Our GPU-accelerated heuristic will be made open-source to further stimulate research on DMMV and its other applications. The code is available at https://anonymous.4open.science/r/AMVM-5F3E/", "date": "2025-08-19", "categories": []} {"id": "2508.14020", "title": "A Biased Random Key Genetic Algorithm for Solving the Longest Run Subsequence Problem", "abstract": "The longest run subsequence (LRS) problem is an NP-hard combinatorial optimization problem belonging to the class of subsequence problems from bioinformatics. In particular, the problem plays a role in genome reassembly. In this paper, we present a solution to the LRS problem using a Biased Random Key Genetic Algorithm (BRKGA). Our approach places particular focus on the computational efficiency of evaluating individuals, which involves converting vectors of gray values into valid solutions to the problem. For comparison purposes, a Max-Min Ant System is developed and implemented. This is in addition to the application of the integer linear programming solver CPLEX for solving all considered problem instances. The computation results show that the proposed BRKGA is currently a state-of-the-art technique for the LRS problem. Nevertheless, the results also show that there is room for improvement, especially in the context of input strings based on large alphabet sizes.", "date": "2025-08-19", "categories": []} {"id": "2409.04653", "title": "Solving Stochastic Orienteering Problems with Chance Constraints Using a GNN Powered Monte Carlo Tree Search", "abstract": "Leveraging the power of a graph neural network (GNN) with message passing, we present a Monte Carlo Tree Search (MCTS) method to solve stochastic orienteering problems with chance constraints. While adhering to an assigned travel budget the algorithm seeks to maximize collected reward while incurring stochastic travel costs. In this context, the acceptable probability of exceeding the assigned budget is expressed as a chance constraint. Our MCTS solution is an online and anytime algorithm alternating planning and execution that determines the next vertex to visit by continuously monitoring the remaining travel budget. The novelty of our work is that the rollout phase in the MCTS framework is implemented using a message passing GNN, predicting both the utility and failure probability of each available action. This allows to enormously expedite the search process. Our experimental evaluation shows that with the proposed method and architecture we manage to efficiently solve complex problem instances while incurring in moderate losses in terms of collected reward. Moreover, we demonstrate how the approach is capable of generalizing beyond the characteristics of the training dataset. The paper's website, open-source code, and supplementary documentation can be found at ucmercedrobotics.github.io/gnn-sop.", "date": "2024-09-06", "categories": []} {"id": "2508.11679", "title": "Lifelong Learner: Discovering Versatile Neural Solvers for Vehicle Routing Problems", "abstract": "Deep learning has been extensively explored to solve vehicle routing problems (VRPs), which yields a range of data-driven neural solvers with promising outcomes. However, most neural solvers are trained to tackle VRP instances in a relatively monotonous context, e.g., simplifying VRPs by using Euclidean distance between nodes and adhering to a single problem size, which harms their off-the-shelf application in different scenarios. To enhance their versatility, this paper presents a novel lifelong learning framework that incrementally trains a neural solver to manage VRPs in distinct contexts. Specifically, we propose a lifelong learner (LL), exploiting a Transformer network as the backbone, to solve a series of VRPs. The inter-context self-attention mechanism is proposed within LL to transfer the knowledge obtained from solving preceding VRPs into the succeeding ones. On top of that, we develop a dynamic context scheduler (DCS), employing the cross-context experience replay to further facilitate LL looking back on the attained policies of solving preceding VRPs. Extensive results on synthetic and benchmark instances (problem sizes up to 18k) show that our LL is capable of discovering effective policies for tackling generic VRPs in varying contexts, which outperforms other neural solvers and achieves the best performance for most VRPs.", "date": "2025-08-08", "categories": []} {"id": "2508.12651", "title": "The Maximum Coverage Model and Recommendation System for UAV Vertiports Location Planning", "abstract": "As urban aerial mobility (UAM) infrastructure development accelerates globally, cities like Shenzhen are planning large-scale vertiport networks (e.g., 1,200+ facilities by 2026). Existing planning frameworks remain inadequate for this complexity due to historical limitations in data granularity and real-world applicability. This paper addresses these gaps by first proposing the Capacitated Dynamic Maximum Covering Location Problem (CDMCLP), a novel optimization framework that simultaneously models urban-scale spatial-temporal demand, heterogeneous user behaviors, and infrastructure capacity constraints. Building on this foundation, we introduce an Integrated Planning Recommendation System that combines CDMCLP with socio-economic factors and dynamic clustering initialization. This system leverages adaptive parameter tuning based on empirical user behavior to generate practical planning solutions. Validation in a Chinese center city demonstrates the effectiveness of the new optimization framework and recommendation system. Under the evaluation and optimization of CDMCLP, the quantitative performance of traditional location methods are exposed and can be improved by 38\\%--52\\%, while the recommendation system shows user-friendliness and the effective integration of complex elements. By integrating mathematical rigor with practical implementation considerations, this hybrid approach bridges the gap between theoretical location modeling and real-world UAM infrastructure planning, offering municipalities a pragmatic tool for vertiport network design.", "date": "2025-08-18", "categories": []} {"id": "2402.02190", "title": "Continuous Parallel Relaxation for Finding Diverse Solutions in Combinatorial Optimization Problems", "abstract": "Finding the optimal solution is often the primary goal in combinatorial optimization (CO). However, real-world applications frequently require diverse solutions rather than a single optimum, particularly in two key scenarios. The first scenario occurs in real-world applications where strictly enforcing every constraint is neither necessary nor desirable. Allowing minor constraint violations can often lead to more cost-effective solutions. This is typically achieved by incorporating the constraints as penalty terms in the objective function, which requires careful tuning of penalty parameters. The second scenario involves cases where CO formulations tend to oversimplify complex real-world factors, such as domain knowledge, implicit trade-offs, or ethical considerations. To address these challenges, generating (i) penalty-diversified solutions by varying penalty intensities and (ii) variation-diversified solutions with distinct structural characteristics provides valuable insights, enabling practitioners to post-select the most suitable solution for their specific needs. However, efficiently discovering these diverse solutions is more challenging than finding a single optimal one. This study introduces Continual Parallel Relaxation Annealing (CPRA), a computationally efficient framework for unsupervised-learning (UL)-based CO solvers that generates diverse solutions within a single training run. CPRA leverages representation learning and parallelization to automatically discover shared representations, substantially accelerating the search for these diverse solutions. Numerical experiments demonstrate that CPRA outperforms existing UL-based solvers in generating these diverse solutions while significantly reducing computational costs.", "date": "2024-02-03", "categories": []} {"id": "2508.10520", "title": "Nonlocal Monte Carlo via Reinforcement Learning", "abstract": "Optimizing or sampling complex cost functions of combinatorial optimization problems is a longstanding challenge across disciplines and applications. When employing family of conventional algorithms based on Markov Chain Monte Carlo (MCMC) such as simulated annealing or parallel tempering, one assumes homogeneous (equilibrium) temperature profiles across input. This instance independent approach was shown to be ineffective for the hardest benchmarks near a computational phase transition when the so-called overlap-gap-property holds. In these regimes conventional MCMC struggles to unfreeze rigid variables, escape suboptimal basins of attraction, and sample high-quality and diverse solutions. In order to mitigate these challenges, Nonequilibrium Nonlocal Monte Carlo (NMC) algorithms were proposed that leverage inhomogeneous temperature profiles thereby accelerating exploration of the configuration space without compromising its exploitation. Here, we employ deep reinforcement learning (RL) to train the nonlocal transition policies of NMC which were previously designed phenomenologically. We demonstrate that the resulting solver can be trained solely by observing energy changes of the configuration space exploration as RL rewards and the local minimum energy landscape geometry as RL states. We further show that the trained policies improve upon the standard MCMC-based and nonlocal simulated annealing on hard uniform random and scale-free random 4-SAT benchmarks in terms of residual energy, time-to-solution, and diversity of solutions metrics.", "date": "2025-08-14", "categories": []} {"id": "2508.09206", "title": "The First Differentiable Transfer-Based Algorithm for Discrete MicroLED Repair", "abstract": "Laser-enabled selective transfer, a key process in high-throughput microLED fabrication, requires computational models that can plan shift sequences to minimize motion of XY stages and adapt to varying optimization objectives across the substrate. We propose the first repair algorithm based on a differentiable transfer module designed to model discrete shifts of transfer platforms, while remaining trainable via gradient-based optimization. Compared to local proximity searching algorithms, our approach achieves superior repair performance and enables more flexible objective designs, such as minimizing the number of steps. Unlike reinforcement learning (RL)-based approaches, our method eliminates the need for handcrafted feature extractors and trains significantly faster, allowing scalability to large arrays. Experiments show a 50% reduction in transfer steps and sub-2-minute planning time on 2000x2000 arrays. This method provides a practical and adaptable solution for accelerating microLED repair in AR/VR and next-generation display fabrication.", "date": "2025-08-10", "categories": []} {"id": "2401.03237", "title": "A New Parallel Cooperative Landscape Smoothing Algorithm and Its Applications on TSP and UBQP", "abstract": "Combinatorial optimization problem (COP) is difficult to solve because of the massive number of local optimal solutions in his solution space. Various methods have been put forward to smooth the solution space of COPs, including homotopic convex (HC) transformation for the traveling salesman problem (TSP). This paper extends the HC transformation approach to unconstrained binary quadratic programming (UBQP) by proposing a method to construct a unimodal toy UBQP of any size. We theoretically prove the unimodality of the constructed toy UBQP. After that, we apply this unimodal toy UBQP to smooth the original UBQP by using the HC transformation framework and empirically verify the smoothing effects. Subsequently, we introduce an iterative algorithmic framework incorporating HC transformation, referred as landscape smoothing iterated local search (LSILS). Our experimental analyses, conducted on various UBQP instances show the effectiveness of LSILS. Furthermore, this paper proposes a parallel cooperative variant of LSILS, denoted as PC-LSILS and apply it to both the UBQP and the TSP. Our experimental findings highlight that PC-LSILS improves the smoothing performance of the HC transformation, and further improves the overall performance of the algorithm.", "date": "2024-01-06", "categories": []} {"id": "2407.19676", "title": "On the Effects of Smoothing Rugged Landscape by Different Toy Problems: A Case Study on UBQP", "abstract": "The hardness of the Unconstrained Binary Quadratic Program (UBQP) problem is due its rugged landscape. Various algorithms have been proposed for UBQP, including the Landscape Smoothing Iterated Local Search (LSILS). Different from other UBQP algorithms, LSILS tries to smooth the rugged landscape by building a convex combination of the original UBQP and a toy UBQP. In this paper, our study further investigates the impact of smoothing rugged landscapes using different toy UBQP problems, including a toy UBQP with matrix ^Q1 (construct by \"+/-1\"), a toy UBQP with matrix ^Q2 (construct by \"+/-i\") and a toy UBQP with matrix ^Q3 (construct randomly). We first assess the landscape flatness of the three toy UBQPs. Subsequently, we test the efficiency of LSILS with different toy UBQPs. Results reveal that the toy UBQP with ^Q1 (construct by \"+/-1\") exhibits the flattest landscape among the three, while the toy UBQP with ^Q3 (construct randomly) presents the most non-flat landscape. Notably, LSILS using the toy UBQP with ^Q2 (construct by \"+/-i\") emerges as the most effective, while ^Q3 (construct randomly) has the poorest result. These findings contribute to a detailed understanding of landscape smoothing techniques in optimizing UBQP.", "date": "2024-07-29", "categories": []} {"id": "2508.08493", "title": "POMO+: Leveraging starting nodes in POMO for solving Capacitated Vehicle Routing Problem", "abstract": "In recent years, reinforcement learning (RL) methods have emerged as a promising approach for solving combinatorial problems. Among RL-based models, POMO has demonstrated strong performance on a variety of tasks, including variants of the Vehicle Routing Problem (VRP). However, there is room for improvement for these tasks. In this work, we improved POMO, creating a method (\\textbf{POMO+}) that leverages the initial nodes to find a solution in a more informed way. We ran experiments on our new model and observed that our solution converges faster and achieves better results. We validated our models on the CVRPLIB dataset and noticed improvements in problem instances with up to 100 customers. We hope that our research in this project can lead to further advancements in the field.", "date": "2025-08-11", "categories": []} {"id": "2508.08659", "title": "Hybrid Node-Destroyer Model with Large Neighborhood Search for Solving the Capacitated Vehicle Routing Problem", "abstract": "In this research, we propose an iterative learning hybrid optimization solver developed to strengthen the performance of metaheuristic algorithms in solving the Capacitated Vehicle Routing Problem (CVRP). The iterative hybrid mechanism integrates the proposed Node-Destroyer Model, a machine learning hybrid model that utilized Graph Neural Networks (GNNs) such identifies and selects customer nodes to guide the Large Neighborhood Search (LNS) operator within the metaheuristic optimization frameworks. This model leverages the structural properties of the problem and solution that can be represented as a graph, to guide strategic selections concerning node removal. The proposed approach reduces operational complexity and scales down the search space involved in the optimization process. The hybrid approach is applied specifically to the CVRP and does not require retraining across problem instances of different sizes. The proposed hybrid mechanism is able to improve the performance of baseline metaheuristic algorithms. Our approach not only enhances the solution quality for standard CVRP benchmarks but also proves scalability on very large-scale instances with up to 30,000 customer nodes. Experimental evaluations on benchmark datasets show that the proposed hybrid mechanism is capable of improving different baseline algorithms, achieving better quality of solutions under similar settings.", "date": "2025-08-12", "categories": []} {"id": "2508.08718", "title": "Generative Modeling for Robust Deep Reinforcement Learning on the Traveling Salesman Problem", "abstract": "The Traveling Salesman Problem (TSP) is a classic NP-hard combinatorial optimization task with numerous practical applications. Classic heuristic solvers can attain near-optimal performance for small problem instances, but become computationally intractable for larger problems. Real-world logistics problems such as dynamically re-routing last-mile deliveries demand a solver with fast inference time, which has led researchers to investigate specialized neural network solvers. However, neural networks struggle to generalize beyond the synthetic data they were trained on. In particular, we show that there exist TSP distributions that are realistic in practice, which also consistently lead to poor worst-case performance for existing neural approaches. To address this issue of distribution robustness, we present Combinatorial Optimization with Generative Sampling (COGS), where training data is sampled from a generative TSP model. We show that COGS provides better data coverage and interpolation in the space of TSP training distributions. We also present TSPLib50, a dataset of realistically distributed TSP samples, which tests real-world generalization ability without conflating this issue with instance size. We evaluate our method on various synthetic datasets as well as TSPLib50, and compare to state-of-the-art neural baselines. We demonstrate that COGS improves distribution robustness, with most performance gains coming from worst-case scenarios.", "date": "2025-08-12", "categories": []} {"id": "2502.12756", "title": "Navigating Demand Uncertainty in Container Shipping: Deep Reinforcement Learning for Enabling Adaptive and Feasible Master Stowage Planning", "abstract": "Reinforcement learning (RL) has shown promise in solving various combinatorial optimization problems. However, conventional RL faces challenges when dealing with complex, real-world constraints, especially when action space feasibility is explicit and dependent on the corresponding state or trajectory. In this work, we address stochastic sequential dynamic decision-making problems with state-dependent constraints. As a relevant and real-world case study, we focus on the master stowage planning problem in container shipping, which aims to optimize revenue and operational costs under demand uncertainty and operational constraints. We propose a deep RL framework with an encoder-decoder model and feasibility layers that satisfy convex constraints and maintain unbiased gradient flow, which embed problem instances, current solutions, and demand uncertainty to guide learning. Experiments show that our model efficiently finds adaptive, feasible solutions that generalize across varying distributions and scale to larger instances, outperforming state-of-the-art baselines in constrained RL and stochastic programming. By uniting artificial intelligence and operations research, our policy empowers humans to make adaptive, uncertainty-aware decisions for resilient and sustainable planning.", "date": "2025-02-18", "categories": []} {"id": "2508.06736", "title": "ParBalans: Parallel Multi-Armed Bandits-based Adaptive Large Neighborhood Search", "abstract": "Solving Mixed-Integer Programming (MIP) problems often requires substantial computational resources due to their combinatorial nature. Parallelization has emerged as a critical strategy to accelerate solution times and enhance scalability to tackle large, complex instances. This paper investigates the parallelization capabilities of Balans, a recently proposed multi-armed bandits-based adaptive large neighborhood search for MIPs. While Balans's modular architecture inherently supports parallel exploration of diverse parameter configurations, this potential has not been thoroughly examined. To address this gap, we introduce ParBalans, an extension that leverages both solver-level and algorithmic-level parallelism to improve performance on challenging MIP instances. Our experimental results demonstrate that ParBalans exhibits competitive performance compared to the state-of-the-art commercial solver Gurobi, particularly on hard optimization benchmarks.", "date": "2025-08-08", "categories": []} {"id": "2411.17080", "title": "DeepMDV: Global Spatial Matching for Multi-depot Vehicle Routing Problems", "abstract": "The rapid growth of online retail and e-commerce has made effective and efficient Vehicle Routing Problem (VRP) solutions essential. To meet rising demand, companies are adding more depots, which changes the VRP problem to a complex optimization task of Multi-Depot VRP (MDVRP) where the routing decisions of vehicles from multiple depots are highly interdependent. The complexities render traditional VRP methods suboptimal and non-scalable for the MDVRP. In this paper, we propose a novel approach to solve MDVRP addressing these interdependencies, hence achieving more effective results. The key idea is, the MDVRP can be broken down into two core spatial tasks: assigning customers to depots and optimizing the sequence of customer visits. We adopt task-decoupling approach and propose a two-stage framework that is scalable: (i) an interdependent partitioning module that embeds spatial and tour context directly into the representation space to globally match customers to depots and assign them to tours; and (ii) an independent routing module that determines the optimal visit sequence within each tour. Extensive experiments on both synthetic and real-world datasets demonstrate that our method outperforms all baselines across varying problem sizes, including the adaptations of learning-based solutions for single-depot VRP. Its adaptability and performance make it a practical and readily deployable solution for real-world logistics challenges.", "date": "2024-11-26", "categories": []} {"id": "2508.00915", "title": "Accelerating Fleet Upgrade Decisions with Machine-Learning Enhanced Optimization", "abstract": "Rental-based business models and increasing sustainability requirements intensify the need for efficient strategies to manage large machine and vehicle fleet renewal and upgrades. Optimized fleet upgrade strategies maximize overall utility, cost, and sustainability. However, conventional fleet optimization does not account for upgrade options and is based on integer programming with exponential runtime scaling, which leads to substantial computational cost when dealing with large fleets and repeated decision-making processes. This contribution firstly suggests an extended integer programming approach that determines optimal renewal and upgrade decisions. The computational burden is addressed by a second, alternative machine learning-based method that transforms the task to a mixed discrete-continuous optimization problem. Both approaches are evaluated in a real-world automotive industry case study, which shows that the machine learning approach achieves near-optimal solutions with significant improvements in the scalability and overall computational performance, thus making it a practical alternative for large-scale fleet management.", "date": "2025-07-29", "categories": []} {"id": "2508.06064", "title": "A Generic Complete Anytime Beam Search for Optimal Decision Tree", "abstract": "Finding an optimal decision tree that minimizes classification error is known to be NP-hard. While exact algorithms based on MILP, CP, SAT, or dynamic programming guarantee optimality, they often suffer from poor anytime behavior -- meaning they struggle to find high-quality decision trees quickly when the search is stopped before completion -- due to unbalanced search space exploration. To address this, several anytime extensions of exact methods have been proposed, such as LDS-DL8.5, Top-k-DL8.5, and Blossom, but they have not been systematically compared, making it difficult to assess their relative effectiveness. In this paper, we propose CA-DL8.5, a generic, complete, and anytime beam search algorithm that extends the DL8.5 framework and unifies some existing anytime strategies. In particular, CA-DL8.5 generalizes previous approaches LDS-DL8.5 and Top-k-DL8.5, by allowing the integration of various heuristics and relaxation mechanisms through a modular design. The algorithm reuses DL8.5's efficient branch-and-bound pruning and trie-based caching, combined with a restart-based beam search that gradually relaxes pruning criteria to improve solution quality over time. Our contributions are twofold: (1) We introduce this new generic framework for exact and anytime decision tree learning, enabling the incorporation of diverse heuristics and search strategies; (2) We conduct a rigorous empirical comparison of several instantiations of CA-DL8.5 -- based on Purity, Gain, Discrepancy, and Top-k heuristics -- using an anytime evaluation metric called the primal gap integral. Experimental results on standard classification benchmarks show that CA-DL8.5 using LDS (limited discrepancy) consistently provides the best anytime performance, outperforming both other CA-DL8.5 variants and the Blossom algorithm while maintaining completeness and optimality guarantees.", "date": "2025-08-08", "categories": []} {"id": "2508.06129", "title": "Study of Robust Features in Formulating Guidance for Heuristic Algorithms for Solving the Vehicle Routing Problem", "abstract": "The Vehicle Routing Problem (VRP) is a complex optimization problem with numerous real-world applications, mostly solved using metaheuristic algorithms due to its $\\mathcal{NP}$-Hard nature. Traditionally, these metaheuristics rely on human-crafted designs developed through empirical studies. However, recent research shows that machine learning methods can be used the structural characteristics of solutions in combinatorial optimization, thereby aiding in designing more efficient algorithms, particularly for solving VRP. Building on this advancement, this study extends the previous research by conducting a sensitivity analysis using multiple classifier models that are capable of predicting the quality of VRP solutions. Hence, by leveraging explainable AI, this research is able to extend the understanding of how these models make decisions. Finally, our findings indicate that while feature importance varies, certain features consistently emerge as strong predictors. Furthermore, we propose a unified framework able of ranking feature impact across different scenarios to illustrate this finding. These insights highlight the potential of feature importance analysis as a foundation for developing a guidance mechanism of metaheuristic algorithms for solving the VRP.", "date": "2025-08-08", "categories": []} {"id": "2503.02992", "title": "RAILGUN: A Unified Convolutional Policy for Multi-Agent Path Finding Across Different Environments and Tasks", "abstract": "Multi-Agent Path Finding (MAPF), which focuses on finding collision-free paths for multiple robots, is crucial for applications ranging from aerial swarms to warehouse automation. Solving MAPF is NP-hard so learning-based approaches for MAPF have gained attention, particularly those leveraging deep neural networks. Nonetheless, despite the community's continued efforts, all learning-based MAPF planners still rely on decentralized planning due to variability in the number of agents and map sizes. We have developed the first centralized learning-based policy for MAPF problem called RAILGUN. RAILGUN is not an agent-based policy but a map-based policy. By leveraging a CNN-based architecture, RAILGUN can generalize across different maps and handle any number of agents. We collect trajectories from rule-based methods to train our model in a supervised way. In experiments, RAILGUN outperforms most baseline methods and demonstrates great zero-shot generalization capabilities on various tasks, maps and agent numbers that were not seen in the training dataset.", "date": "2025-03-04", "categories": []} {"id": "2503.03779", "title": "Accelerating Focal Search in Multi-Agent Path Finding with Tighter Lower Bounds", "abstract": "Multi-Agent Path Finding (MAPF) involves finding collision-free paths for multiple agents while minimizing a cost function--an NP-hard problem. Bounded suboptimal methods like Enhanced Conflict-Based Search (ECBS) and Explicit Estimation CBS (EECBS) balance solution quality with computational efficiency using focal search mechanisms. While effective, traditional focal search faces a limitation: the lower bound (LB) value determining which nodes enter the FOCAL list often increases slowly in early search stages, resulting in a constrained search space that delays finding valid solutions. In this paper, we propose a novel bounded suboptimal algorithm, double-ECBS (DECBS), to address this issue by first determining the maximum LB value and then employing a best-first search guided by this LB to find a collision-free path. Experimental results demonstrate that DECBS outperforms ECBS in most test cases and is compatible with existing optimization techniques. DECBS can reduce nearly 30% high-level CT nodes and 50% low-level focal search nodes. When agent density is moderate to high, DECBS achieves a 23.5% average runtime improvement over ECBS with identical suboptimality bounds and optimizations.", "date": "2025-03-04", "categories": []} {"id": "2505.09742", "title": "A Generative Neural Annealer for Black-Box Combinatorial Optimization", "abstract": "We propose a generative, end-to-end solver for black-box combinatorial optimization that emphasizes both sample efficiency and solution quality on NP problems. Drawing inspiration from annealing-based algorithms, we treat the black-box objective as an energy function and train a neural network to model the associated Boltzmann distribution. By conditioning on temperature, the network captures a continuum of distributions--from near-uniform at high temperatures to sharply peaked around global optima at low temperatures--thereby learning the structure of the energy landscape and facilitating global optimization. When queries are expensive, the temperature-dependent distributions naturally enable data augmentation and improve sample efficiency. When queries are cheap but the problem remains hard, the model learns implicit variable interactions, effectively \"opening\" the black box. We validate our approach on challenging combinatorial tasks under both limited and unlimited query budgets, showing competitive performance against state-of-the-art black-box optimizers.", "date": "2025-05-14", "categories": []} {"id": "2508.03777", "title": "When Agents Break Down in Multiagent Path Finding", "abstract": "In Multiagent Path Finding (MAPF), the goal is to compute efficient, collision-free paths for multiple agents navigating a network from their sources to targets, minimizing the schedule's makespan-the total time until all agents reach their destinations. We introduce a new variant that formally models scenarios where some agents may experience delays due to malfunctions, posing significant challenges for maintaining optimal schedules.\n Recomputing an entirely new schedule from scratch after each malfunction is often computationally infeasible. To address this, we propose a framework for dynamic schedule adaptation that does not rely on full replanning. Instead, we develop protocols enabling agents to locally coordinate and adjust their paths on the fly. We prove that following our primary communication protocol, the increase in makespan after k malfunctions is bounded by k additional turns, effectively limiting the impact of malfunctions on overall efficiency. Moreover, recognizing that agents may have limited computational capabilities, we also present a secondary protocol that shifts the necessary computations onto the network's nodes, ensuring robustness without requiring enhanced agent processing power. Our results demonstrate that these protocols provide a practical, scalable approach to resilient multiagent navigation in the face of agent failures.", "date": "2025-08-05", "categories": []} {"id": "2508.04235", "title": "Circuit-Aware SAT Solving: Guiding CDCL via Conditional Probabilities", "abstract": "Circuit Satisfiability (CSAT) plays a pivotal role in Electronic Design Automation. The standard workflow for solving CSAT problems converts circuits into Conjunctive Normal Form (CNF) and employs generic SAT solvers powered by Conflict-Driven Clause Learning (CDCL). However, this process inherently discards rich structural and functional information, leading to suboptimal solver performance. To address this limitation, we introduce CASCAD, a novel circuit-aware SAT solving framework that directly leverages circuit-level conditional probabilities computed via Graph Neural Networks (GNNs). By explicitly modeling gate-level conditional probabilities, CASCAD dynamically guides two critical CDCL heuristics -- variable phase selection and clause managementto significantly enhance solver efficiency. Extensive evaluations on challenging real-world Logical Equivalence Checking (LEC) benchmarks demonstrate that CASCAD reduces solving times by up to 10x compared to state-of-the-art CNF-based approaches, achieving an additional 23.5% runtime reduction via our probability-guided clause filtering strategy. Our results underscore the importance of preserving circuit-level structural insights within SAT solvers, providing a robust foundation for future improvements in SAT-solving efficiency and EDA tool design.", "date": "2025-08-06", "categories": []} {"id": "2508.03406", "title": "Multi-Objective Infeasibility Diagnosis for Routing Problems Using Large Language Models", "abstract": "In real-world routing problems, users often propose conflicting or unreasonable requirements, which result in infeasible optimization models due to overly restrictive or contradictory constraints, leading to an empty feasible solution set. Existing Large Language Model (LLM)-based methods attempt to diagnose infeasible models, but modifying such models often involves multiple potential adjustments that these methods do not consider. To fill this gap, we introduce Multi-Objective Infeasibility Diagnosis (MOID), which combines LLM agents and multi-objective optimization within an automatic routing solver, to provide a set of representative actionable suggestions. Specifically, MOID employs multi-objective optimization to consider both path cost and constraint violation, generating a set of trade-off solutions, each encompassing varying degrees of model adjustments. To extract practical insights from these solutions, MOID utilizes LLM agents to generate a solution analysis function for the infeasible model. This function analyzes these distinct solutions to diagnose the original infeasible model, providing users with diverse diagnostic insights and suggestions. Finally, we compare MOID with several LLM-based methods on 50 types of infeasible routing problems. The results indicate that MOID automatically generates multiple diagnostic suggestions in a single run, providing more practical insights for restoring model feasibility and decision-making compared to existing methods.", "date": "2025-08-05", "categories": []} {"id": "2508.00887", "title": "FRAM: Frobenius-Regularized Assignment Matching with Mixed-Precision Computing", "abstract": "Graph matching, typically formulated as a Quadratic Assignment Problem (QAP), seeks to establish node correspondences between two graphs. To address the NP-hardness of QAP, some existing methods adopt projection-based relaxations that embed the problem into the convex hull of the discrete domain. However, these relaxations inevitably enlarge the feasible set, introducing two sources of error: numerical scale sensitivity and geometric misalignment between the relaxed and original domains. To alleviate these errors, we propose a novel relaxation framework by reformulating the projection step as a Frobenius-regularized Linear Assignment (FRA) problem, where a tunable regularization term mitigates feasible region inflation. This formulation enables normalization-based operations to preserve numerical scale invariance without compromising accuracy. To efficiently solve FRA, we propose the Scaling Doubly Stochastic Normalization (SDSN) algorithm. Building on its favorable computational properties, we develop a theoretically grounded mixed-precision architecture to achieve substantial acceleration. Comprehensive CPU-based benchmarks demonstrate that FRAM consistently outperforms all baseline methods under identical precision settings. When combined with a GPU-based mixed-precision architecture, FRAM achieves up to 370X speedup over its CPU-FP64 counterpart, with negligible loss in solution accuracy.", "date": "2025-07-26", "categories": []} {"id": "2508.01476", "title": "CARGO: A Co-Optimization Framework for EV Charging and Routing in Goods Delivery Logistics", "abstract": "With growing interest in sustainable logistics, electric vehicle (EV)-based deliveries offer a promising alternative for urban distribution. However, EVs face challenges due to their limited battery capacity, requiring careful planning for recharging. This depends on factors such as the charging point (CP) availability, cost, proximity, and vehicles' state of charge (SoC). We propose CARGO, a framework addressing the EV-based delivery route planning problem (EDRP), which jointly optimizes route planning and charging for deliveries within time windows. After proving the problem's NP-hardness, we propose a mixed integer linear programming (MILP)-based exact solution and a computationally efficient heuristic method. Using real-world datasets, we evaluate our methods by comparing the heuristic to the MILP solution, and benchmarking it against baseline strategies, Earliest Deadline First (EDF) and Nearest Delivery First (NDF). The results show up to 39% and 22% reductions in the charging cost over EDF and NDF, respectively, while completing comparable deliveries.", "date": "2025-08-02", "categories": []} {"id": "2508.02510", "title": "On Distributional Dependent Performance of Classical and Neural Routing Solvers", "abstract": "Neural Combinatorial Optimization aims to learn to solve a class of combinatorial problems through data-driven methods and notably through employing neural networks by learning the underlying distribution of problem instances. While, so far neural methods struggle to outperform highly engineered problem specific meta-heuristics, this work explores a novel approach to formulate the distribution of problem instances to learn from and, more importantly, plant a structure in the sampled problem instances. In application to routing problems, we generate large problem instances that represent custom base problem instance distributions from which training instances are sampled. The test instances to evaluate the methods on the routing task consist of unseen problems sampled from the underlying large problem instance. We evaluate representative NCO methods and specialized Operation Research meta heuristics on this novel task and demonstrate that the performance gap between neural routing solvers and highly specialized meta-heuristics decreases when learning from sub-samples drawn from a fixed base node distribution.", "date": "2025-08-04", "categories": []} {"id": "2506.14781", "title": "Two-dimensional Parallel Tempering for Constrained Optimization", "abstract": "Sampling Boltzmann probability distributions plays a key role in machine learning and optimization, motivating the design of hardware accelerators such as Ising machines. While the Ising model can in principle encode arbitrary optimization problems, practical implementations are often hindered by soft constraints that either slow down mixing when too strong, or fail to enforce feasibility when too weak. We introduce a two-dimensional extension of the powerful parallel tempering algorithm (PT) that addresses this challenge by adding a second dimension of replicas interpolating the penalty strengths. This scheme ensures constraint satisfaction in the final replicas, analogous to low-energy states at low temperature. The resulting two-dimensional parallel tempering algorithm (2D-PT) improves mixing in heavily constrained replicas and eliminates the need to explicitly tune the penalty strength. In a representative example of graph sparsification with copy constraints, 2D-PT achieves near-ideal mixing, with Kullback-Leibler divergence decaying as O(1/t). When applied to sparsified Wishart instances, 2D-PT yields orders of magnitude speedup over conventional PT with the same number of replicas. The method applies broadly to constrained Ising problems and can be deployed on existing Ising machines.", "date": "2025-05-24", "categories": []} {"id": "2501.03518", "title": "Transfer Learning for Deep-Unfolded Combinatorial Optimization Solver with Quantum Annealer", "abstract": "Quantum annealing (QA) has attracted research interest as a sampler and combinatorial optimization problem (COP) solver. A recently proposed sampling-based solver for QA significantly reduces the required number of qubits, being capable of large COPs. In relation to this, a trainable sampling-based COP solver has been proposed that optimizes its internal parameters from a dataset by using a deep learning technique called deep unfolding. Although learning the internal parameters accelerates the convergence speed, the sampler in the trainable solver is restricted to using a classical sampler owing to the training cost. In this study, to utilize QA in the trainable solver, we propose classical-quantum transfer learning, where parameters are trained classically, and the trained parameters are used in the solver with QA. The results of numerical experiments demonstrate that the trainable quantum COP solver using classical-quantum transfer learning improves convergence speed and execution time over the original solver.", "date": "2025-01-07", "categories": []} {"id": "2507.22338", "title": "Parametrized Multi-Agent Routing via Deep Attention Models", "abstract": "We propose a scalable deep learning framework for parametrized sequential decision-making (ParaSDM), where multiple agents jointly optimize discrete action policies and shared continuous parameters. A key subclass of this setting arises in Facility-Location and Path Optimization (FLPO), where multi-agent systems must simultaneously determine optimal routes and facility locations, aiming to minimize the cumulative transportation cost within the network. FLPO problems are NP-hard due to their mixed discrete-continuous structure and highly non-convex objective. To address this, we integrate the Maximum Entropy Principle (MEP) with a neural policy model called the Shortest Path Network (SPN)-a permutation-invariant encoder-decoder that approximates the MEP solution while enabling efficient gradient-based optimization over shared parameters. The SPN achieves up to 100$\\times$ speedup in policy inference and gradient computation compared to MEP baselines, with an average optimality gap of approximately 6% across a wide range of problem sizes. Our FLPO approach yields over 10$\\times$ lower cost than metaheuristic baselines while running significantly faster, and matches Gurobi's optimal cost with annealing at a 1500$\\times$ speedup-establishing a new state of the art for ParaSDM problems. These results highlight the power of structured deep models for solving large-scale mixed-integer optimization tasks.", "date": "2025-07-30", "categories": []} {"id": "2507.22440", "title": "Nearest-Better Network for Visualizing and Analyzing Combinatorial Optimization Problems: A Unified Tool", "abstract": "The Nearest-Better Network (NBN) is a powerful method to visualize sampled data for continuous optimization problems while preserving multiple landscape features. However, the calculation of NBN is very time-consuming, and the extension of the method to combinatorial optimization problems is challenging but very important for analyzing the algorithm's behavior. This paper provides a straightforward theoretical derivation showing that the NBN network essentially functions as the maximum probability transition network for algorithms. This paper also presents an efficient NBN computation method with logarithmic linear time complexity to address the time-consuming issue. By applying this efficient NBN algorithm to the OneMax problem and the Traveling Salesman Problem (TSP), we have made several remarkable discoveries for the first time: The fitness landscape of OneMax exhibits neutrality, ruggedness, and modality features. The primary challenges of TSP problems are ruggedness, modality, and deception. Two state-of-the-art TSP algorithms (i.e., EAX and LKH) have limitations when addressing challenges related to modality and deception, respectively. LKH, based on local search operators, fails when there are deceptive solutions near global optima. EAX, which is based on a single population, can efficiently maintain diversity. However, when multiple attraction basins exist, EAX retains individuals within multiple basins simultaneously, reducing inter-basin interaction efficiency and leading to algorithm's stagnation.", "date": "2025-07-30", "categories": []} {"id": "2506.20031", "title": "Automated Generation of Diverse Courses of Actions for Multi-Agent Operations using Binary Optimization and Graph Learning", "abstract": "Operations in disaster response, search \\& rescue, and military missions that involve multiple agents demand automated processes to support the planning of the courses of action (COA). Moreover, traverse-affecting changes in the environment (rain, snow, blockades, etc.) may impact the expected performance of a COA, making it desirable to have a pool of COAs that are diverse in task distributions across agents. Further, variations in agent capabilities, which could be human crews and/or autonomous systems, present practical opportunities and computational challenges to the planning process. This paper presents a new theoretical formulation and computational framework to generate such diverse pools of COAs for operations with soft variations in agent-task compatibility. Key to the problem formulation is a graph abstraction of the task space and the pool of COAs itself to quantify its diversity. Formulating the COAs as a centralized multi-robot task allocation problem, a genetic algorithm is used for (order-ignoring) allocations of tasks to each agent that jointly maximize diversity within the COA pool and overall compatibility of the agent-task mappings. A graph neural network is trained using a policy gradient approach to then perform single agent task sequencing in each COA, which maximizes completion rates adaptive to task features. Our tests of the COA generation process in a simulated environment demonstrate significant performance gain over a random walk baseline, small optimality gap in task sequencing, and execution time of about 50 minutes to plan up to 20 COAs for 5 agent/100 task operations.", "date": "2025-06-24", "categories": []} {"id": "2507.21963", "title": "SLA-Centric Automated Algorithm Selection Framework for Cloud Environments", "abstract": "Cloud computing offers on-demand resource access, regulated by Service-Level Agreements (SLAs) between consumers and Cloud Service Providers (CSPs). SLA violations can impact efficiency and CSP profitability. In this work, we propose an SLA-aware automated algorithm-selection framework for combinatorial optimization problems in resource-constrained cloud environments. The framework uses an ensemble of machine learning models to predict performance and rank algorithm-hardware pairs based on SLA constraints. We also apply our framework to the 0-1 knapsack problem. We curate a dataset comprising instance specific features along with memory usage, runtime, and optimality gap for 6 algorithms. As an empirical benchmark, we evaluate the framework on both classification and regression tasks. Our ablation study explores the impact of hyperparameters, learning approaches, and large language models effectiveness in regression, and SHAP-based interpretability.", "date": "2025-07-29", "categories": []} {"id": "2504.06091", "title": "Real-Time LaCAM for Real-Time MAPF", "abstract": "The vast majority of Multi-Agent Path Finding (MAPF) methods with completeness guarantees require planning full-horizon paths. However, planning full-horizon paths can take too long and be impractical in real-world applications. Instead, real-time planning and execution, which only allows the planner a finite amount of time before executing and replanning, is more practical for real-world multi-agent systems. Several methods utilize real-time planning schemes but none are provably complete, which leads to livelock or deadlock. Our main contribution is Real-Time LaCAM, the first Real-Time MAPF method with provable completeness guarantees. We do this by leveraging LaCAM (Okumura 2023) in an incremental fashion. Our results show how we can iteratively plan for congested environments with a cutoff time of milliseconds while still maintaining the same success rate as full-horizon LaCAM. We also show how it can be used with a single-step learned MAPF policy.", "date": "2025-04-08", "categories": []} {"id": "2507.20010", "title": "Finding Personalized Good-Enough Solutions to Unsatisfiable Stable Roommates Problems", "abstract": "The Stable Roommates problems are characterized by the preferences of agents over other agents as roommates. A solution is a partition of the agents into pairs that are acceptable to each other (i.e., they are in the preference lists of each other), and the matching is stable (i.e., there do not exist any two agents who prefer each other to their roommates, and thus block the matching). Motivated by real-world applications, and considering that stable roommates problems do not always have solutions, we continue our studies to compute \"good-enough\" matchings. In addition to the agents' habits and habitual preferences, we consider their networks of preferred friends, and introduce a method to generate personalized solutions to stable roommates problems. We illustrate the usefulness of our method with examples and empirical evaluations.", "date": "2025-07-26", "categories": []} {"id": "2412.14382", "title": "Balans: Multi-Armed Bandits-based Adaptive Large Neighborhood Search for Mixed-Integer Programming Problem", "abstract": "Mixed-integer programming (MIP) is a powerful paradigm for modeling and solving various important combinatorial optimization problems. Recently, learning-based approaches have shown a potential to speed up MIP solving via offline training that then guides important design decisions during the search. However, a significant drawback of these methods is their heavy reliance on offline training, which requires collecting training datasets and computationally costly training epochs yet offering only limited generalization to unseen (larger) instances. In this paper, we propose Balans, an adaptive meta-solver for MIPs with online learning capability that does not require any supervision or apriori training. At its core, Balans is based on adaptive large-neighborhood search, operating on top of an MIP solver by successive applications of destroy and repair neighborhood operators. During the search, the selection among different neighborhood definitions is guided on the fly for the instance at hand via multi-armed bandit algorithms. Our extensive experiments on hard optimization instances show that Balans offers significant performance gains over the default MIP solver, is better than committing to any single best neighborhood, and improves over the state-of-the-art large-neighborhood search for MIPs. Finally, we release Balans as a highly configurable, MIP solver agnostic, open-source software.", "date": "2024-12-18", "categories": []} {"id": "2507.17054", "title": "New Mechanisms in Flex Distribution for Bounded Suboptimal Multi-Agent Path Finding", "abstract": "Multi-Agent Path Finding (MAPF) is the problem of finding a set of collision-free paths, one for each agent in a shared environment. Its objective is to minimize the sum of path costs (SOC), where the path cost of each agent is defined as the travel time from its start location to its target location. Explicit Estimation Conflict-Based Search (EECBS) is the leading algorithm for bounded-suboptimal MAPF, with the SOC of the solution being at most a user-specified factor $w$ away from optimal. EECBS maintains sets of paths and a lower bound $LB$ on the optimal SOC. Then, it iteratively selects a set of paths whose SOC is at most $w \\cdot LB$ and introduces constraints to resolve collisions. For each path in a set, EECBS maintains a lower bound on its optimal path that satisfies constraints. By finding an individually bounded-suboptimal path with cost at most a threshold of $w$ times its lower bound, EECBS guarantees to find a bounded-suboptimal solution. To speed up EECBS, previous work uses flex distribution to increase the threshold. Though EECBS with flex distribution guarantees to find a bounded-suboptimal solution, increasing the thresholds may push the SOC beyond $w \\cdot LB$, forcing EECBS to switch among different sets of paths instead of resolving collisions on a particular set of paths, and thus reducing efficiency. To address this issue, we propose Conflict-Based Flex Distribution that distributes flex in proportion to the number of collisions. We also estimate the delays needed to satisfy constraints and propose Delay-Based Flex Distribution. On top of that, we propose Mixed-Strategy Flex Distribution, combining both in a hierarchical framework. We prove that EECBS with our new flex distribution mechanisms is complete and bounded-suboptimal. Our experiments show that our approaches outperform the original (greedy) flex distribution.", "date": "2025-07-22", "categories": []} {"id": "2507.17063", "title": "Compatibility of Max and Sum Objectives for Committee Selection and $k$-Facility Location", "abstract": "We study a version of the metric facility location problem (or, equivalently, variants of the committee selection problem) in which we must choose $k$ facilities in an arbitrary metric space to serve some set of clients $C$. We consider four different objectives, where each client $i\\in C$ attempts to minimize either the sum or the maximum of its distance to the chosen facilities, and where the overall objective either considers the sum or the maximum of the individual client costs. Rather than optimizing a single objective at a time, we study how compatible these objectives are with each other, and show the existence of solutions which are simultaneously close-to-optimum for any pair of the above objectives. Our results show that when choosing a set of facilities or a representative committee, it is often possible to form a solution which is good for several objectives at the same time, instead of sacrificing one desideratum to achieve another.", "date": "2025-07-22", "categories": []} {"id": "2507.01225", "title": "Capacity Planning and Scheduling for Jobs with Uncertainty in Resource Usage and Duration", "abstract": "Organizations around the world schedule jobs (programs) regularly to perform various tasks dictated by their end users. With the major movement towards using a cloud computing infrastructure, our organization follows a hybrid approach with both cloud and on-prem servers. The objective of this work is to perform capacity planning, i.e., estimate resource requirements, and job scheduling for on-prem grid computing environments. A key contribution of our approach is handling uncertainty in both resource usage and duration of the jobs, a critical aspect in the finance industry where stochastic market conditions significantly influence job characteristics. For capacity planning and scheduling, we simultaneously balance two conflicting objectives: (a) minimize resource usage, and (b) provide high quality-of-service to the end users by completing jobs by their requested deadlines. We propose approximate approaches using deterministic estimators and pair sampling-based constraint programming. Our best approach (pair sampling-based) achieves much lower peak resource usage compared to manual scheduling without compromising on the quality-of-service.", "date": "2025-07-01", "categories": []} {"id": "2507.15866", "title": "Purchase and Production Optimization in a Meat Processing Plant", "abstract": "The food production industry, especially the meat production sector, faces many challenges that have even escalated due to the recent outbreak of the energy crisis in the European Union. Therefore, efficient use of input materials is an essential aspect affecting the profit of such companies. This paper addresses an optimization problem concerning the purchase and subsequent material processing we solved for a meat processing company. Unlike the majority of existing papers, we do not concentrate on how this problem concerns supply chain management, but we focus purely on the production stage. The problem involves the concept of alternative ways of material processing, stock of material with different expiration dates, and extra constraints widely neglected in the current literature, namely, the minimum order quantity and the minimum percentage in alternatives. We prove that each of these two constraints makes the problem \\mbox{$\\mathcal{NP}$-hard}, and hence we design a simple iterative approach based on integer linear programming that allows us to solve real-life instances even using an open-source integer linear programming solver. Another advantage of this approach is that it mitigates numerical issues, caused by the extensive range of data values, we experienced with a commercial solver. The results obtained using real data from the meat processing company showed that our algorithm can find the optimum solution in a few seconds for all considered use cases.", "date": "2025-07-14", "categories": []} {"id": "2306.17100", "title": "RL4CO: an Extensive Reinforcement Learning for Combinatorial Optimization Benchmark", "abstract": "Combinatorial optimization (CO) is fundamental to several real-world applications, from logistics and scheduling to hardware design and resource allocation. Deep reinforcement learning (RL) has recently shown significant benefits in solving CO problems, reducing reliance on domain expertise and improving computational efficiency. However, the absence of a unified benchmarking framework leads to inconsistent evaluations, limits reproducibility, and increases engineering overhead, raising barriers to adoption for new researchers. To address these challenges, we introduce RL4CO, a unified and extensive benchmark with in-depth library coverage of 27 CO problem environments and 23 state-of-the-art baselines. Built on efficient software libraries and best practices in implementation, RL4CO features modularized implementation and flexible configurations of diverse environments, policy architectures, RL algorithms, and utilities with extensive documentation. RL4CO helps researchers build on existing successes while exploring and developing their own designs, facilitating the entire research process by decoupling science from heavy engineering. We finally provide extensive benchmark studies to inspire new insights and future work. RL4CO has already attracted numerous researchers in the community and is open-sourced at https://github.com/ai4co/rl4co.", "date": "2023-06-29", "categories": []} {"id": "2503.10968", "title": "Combinatorial Optimization for All: Using LLMs to Aid Non-Experts in Improving Optimization Algorithms", "abstract": "Large Language Models (LLMs) have shown notable potential in code generation for optimization algorithms, unlocking exciting new opportunities. This paper examines how LLMs, rather than creating algorithms from scratch, can improve existing ones without the need for specialized expertise. To explore this potential, we selected 10 baseline optimization algorithms from various domains (metaheuristics, reinforcement learning, deterministic, and exact methods) to solve the classic Travelling Salesman Problem. The results show that our simple methodology often results in LLM-generated algorithm variants that improve over the baseline algorithms in terms of solution quality, reduction in computational time, and simplification of code complexity, all without requiring specialized optimization knowledge or advanced algorithmic implementation skills.", "date": "2025-03-14", "categories": []} {"id": "2505.06178", "title": "A Large Language Model-Enhanced Q-learning for Capacitated Vehicle Routing Problem with Time Windows", "abstract": "The Capacitated Vehicle Routing Problem with Time Windows (CVRPTW) is a classic NP-hard combinatorial optimization problem widely applied in logistics distribution and transportation management. Its complexity stems from the constraints of vehicle capacity and time windows, which pose significant challenges to traditional approaches. Advances in Large Language Models (LLMs) provide new possibilities for finding approximate solutions to CVRPTW. This paper proposes a novel LLM-enhanced Q-learning framework to address the CVRPTW with real-time emergency constraints. Our solution introduces an adaptive two-phase training mechanism that transitions from the LLM-guided exploration phase to the autonomous optimization phase of Q-network. To ensure reliability, we design a three-tier self-correction mechanism based on the Chain-of-Thought (CoT) for LLMs: syntactic validation, semantic verification, and physical constraint enforcement. In addition, we also prioritized replay of the experience generated by LLMs to amplify the regulatory role of LLMs in the architecture. Experimental results demonstrate that our framework achieves a 7.3\\% average reduction in cost compared to traditional Q-learning, with fewer training steps required for convergence.", "date": "2025-05-09", "categories": []} {"id": "2507.15063", "title": "Quantum Annealing for Machine Learning: Applications in Feature Selection, Instance Selection, and Clustering", "abstract": "This paper explores the applications of quantum annealing (QA) and classical simulated annealing (SA) to a suite of combinatorial optimization problems in machine learning, namely feature selection, instance selection, and clustering. We formulate each task as a Quadratic Unconstrained Binary Optimization (QUBO) problem and implement both quantum and classical solvers to compare their effectiveness. For feature selection, we propose several QUBO configurations that balance feature importance and redundancy, showing that quantum annealing (QA) produces solutions that are computationally more efficient. In instance selection, we propose a few novel heuristics for instance-level importance measures that extend existing methods. For clustering, we embed a classical-to-quantum pipeline, using classical clustering followed by QUBO-based medoid refinement, and demonstrate consistent improvements in cluster compactness and retrieval metrics. Our results suggest that QA can be a competitive and efficient tool for discrete machine learning optimization, even within the constraints of current quantum hardware.", "date": "2025-07-20", "categories": []} {"id": "2507.10397", "title": "Instance space analysis of the capacitated vehicle routing problem", "abstract": "This paper seeks to advance CVRP research by addressing the challenge of understanding the nuanced relationships between instance characteristics and metaheuristic (MH) performance. We present Instance Space Analysis (ISA) as a valuable tool that allows for a new perspective on the field. By combining the ISA methodology with a dataset from the DIMACS 12th Implementation Challenge on Vehicle Routing, our research enabled the identification of 23 relevant instance characteristics. Our use of the PRELIM, SIFTED, and PILOT stages, which employ dimensionality reduction and machine learning methods, allowed us to create a two-dimensional projection of the instance space to understand how the structure of instances affect the behavior of MHs. A key contribution of our work is that we provide a projection matrix, which makes it straightforward to incorporate new instances into this analysis and allows for a new method for instance analysis in the CVRP field.", "date": "2025-07-14", "categories": []} {"id": "2410.06187", "title": "A column generation algorithm with dynamic constraint aggregation for minimum sum-of-squares clustering", "abstract": "The minimum sum-of-squares clustering problem (MSSC), also known as $k$-means clustering, refers to the problem of partitioning $n$ data points into $k$ clusters, with the objective of minimizing the total sum of squared Euclidean distances between each point and the center of its assigned cluster. We propose an efficient algorithm for solving large-scale MSSC instances, which combines column generation (CG) with dynamic constraint aggregation (DCA) to effectively reduce the number of constraints considered in the CG master problem. DCA was originally conceived to reduce degeneracy in set partitioning problems by utilizing an aggregated restricted master problem obtained from a partition of the set partitioning constraints into disjoint clusters. In this work, we explore the use of DCA within a CG algorithm for MSSC exact solution. Our method is fine-tuned by a series of ablation studies on DCA design choices, and is demonstrated to significantly outperform existing state-of-the-art exact approaches available in the literature.", "date": "2024-10-08", "categories": []} {"id": "2504.10733", "title": "Cross-Problem Parameter Transfer in Quantum Approximate Optimization Algorithm: A Machine Learning Approach", "abstract": "Quantum Approximate Optimization Algorithm (QAOA) is one of the most promising candidates to achieve the quantum advantage in solving combinatorial optimization problems. The process of finding a good set of variational parameters in the QAOA circuit has proven to be challenging due to multiple factors, such as barren plateaus. As a result, there is growing interest in exploiting parameter transferability, where parameter sets optimized for one problem instance are transferred to another that could be more complex either to estimate the solution or to serve as a warm start for further optimization. But can we transfer parameters from one class of problems to another? Leveraging parameter sets learned from a well-studied class of problems could help navigate the less studied one, reducing optimization overhead and mitigating performance pitfalls. In this paper, we study whether pretrained QAOA parameters of MaxCut can be used as is or to warm start the Maximum Independent Set (MIS) circuits. Specifically, we design machine learning models to find good donor candidates optimized on MaxCut and apply their parameters to MIS acceptors. Our experimental results show that such parameter transfer can significantly reduce the number of optimization iterations required while achieving comparable approximation ratios.", "date": "2025-04-14", "categories": []} {"id": "2507.08053", "title": "Tree-Structured Parzen Estimator Can Solve Black-Box Combinatorial Optimization More Efficiently", "abstract": "Tree-structured Parzen estimator (TPE) is a versatile hyperparameter optimization (HPO) method supported by popular HPO tools. Since these HPO tools have been developed in line with the trend of deep learning (DL), the problem setups often used in the DL domain have been discussed for TPE such as multi-objective optimization and multi-fidelity optimization. However, the practical applications of HPO are not limited to DL, and black-box combinatorial optimization is actively utilized in some domains, e.g., chemistry and biology. As combinatorial optimization has been an untouched, yet very important, topic in TPE, we propose an efficient combinatorial optimization algorithm for TPE. In this paper, we first generalize the categorical kernel with the numerical kernel in TPE, enabling us to introduce a distance structure to the categorical kernel. Then we discuss modifications for the newly developed kernel to handle a large combinatorial search space. These modifications reduce the time complexity of the kernel calculation with respect to the size of a combinatorial search space. In the experiments using synthetic problems, we verified that our proposed method identifies better solutions with fewer evaluations than the original TPE. Our algorithm is available in Optuna, an open-source framework for HPO.", "date": "2025-07-10", "categories": []} {"id": "2507.10636", "title": "GeoHopNet: Hopfield-Augmented Sparse Spatial Attention for Dynamic UAV Site Location Problem", "abstract": "The rapid development of urban low-altitude unmanned aerial vehicle (UAV) economy poses new challenges for dynamic site selection of UAV landing points and supply stations. Traditional deep reinforcement learning methods face computational complexity bottlenecks, particularly with standard attention mechanisms, when handling large-scale urban-level location problems. This paper proposes GeoHopNet, a Hopfield-augmented sparse spatial attention network specifically designed for dynamic UAV site location problems. Our approach introduces four core innovations: (1) distance-biased multi-head attention mechanism that explicitly encodes spatial geometric information; (2) K-nearest neighbor sparse attention that reduces computational complexity from $O(N^2)$ to $O(NK)$; (3) a modern Hopfield external memory module; and (4) a memory regularization strategy. Experimental results demonstrate that GeoHopNet extends the boundary of solvable problem sizes. For large-scale instances with 1,000 nodes, where standard attention models become prohibitively slow (over 3 seconds per instance) and traditional solvers fail, GeoHopNet finds high-quality solutions (0.22\\% optimality gap) in under 0.1 seconds. Compared to the state-of-the-art ADNet baseline on 100-node instances, our method improves solution quality by 22.2\\% and is 1.8$\\times$ faster.", "date": "2025-07-14", "categories": []} {"id": "2401.03692", "title": "Boosting Column Generation with Graph Neural Networks for Joint Rider Trip Planning and Crew Shift Scheduling", "abstract": "Optimizing service schedules is pivotal to the reliable, efficient, and inclusive on-demand mobility. This pressing challenge is further exacerbated by the increasing needs of an aging population, the oversubscription of existing services, and the lack of effective solution methods. This study addresses the intricacies of service scheduling, by jointly optimizing rider trip planning and crew scheduling for a complex dynamic mobility service. The resulting optimization problems are extremely challenging computationally for state-of-the-art methods. To address this fundamental gap, this paper introduces the Joint Rider Trip Planning and Crew Shift Scheduling Problem (JRTPCSSP) and a novel solution method, called Attention and Gated GNN-Informed Column Generation (AGGNNI-CG), that hybridizes column generation and machine learning to obtain near-optimal solutions to the JRTPCSSP with real-life constraints of the application. The key idea of the machine-learning component is to dramatically reduce the number of paths to explore in the pricing problem, accelerating the most time-consuming component of the column generation. The machine learning component is a graph neural network with an attention mechanism and a gated architecture, which is particularly suited to cater for the different input sizes coming from daily operations. AGGNNI-CG has been applied to a challenging, real-world dataset from the Paratransit system of Chatham County in Georgia. It produces substantial improvements compared to the baseline column generation approach, which typically cannot produce high-quality feasible solutions in reasonable time on large-scale complex instances. AGGNNI-CG also produces significant improvements in service quality compared to the existing system.", "date": "2024-01-08", "categories": []} {"id": "2507.07159", "title": "Large-scale portfolio optimization with variational neural annealing", "abstract": "Portfolio optimization is a routine asset management operation conducted in financial institutions around the world. However, under real-world constraints such as turnover limits and transaction costs, its formulation becomes a mixed-integer nonlinear program that current mixed-integer optimizers often struggle to solve. We propose mapping this problem onto a classical Ising-like Hamiltonian and solving it with Variational Neural Annealing (VNA), via its classical formulation implemented using autoregressive neural networks. We demonstrate that VNA can identify near-optimal solutions for portfolios comprising more than 2,000 assets and yields performance comparable to that of state-of-the-art optimizers, such as Mosek, while exhibiting faster convergence on hard instances. Finally, we present a dynamical finite-size scaling analysis applied to the S&P 500, Russell 1000, and Russell 3000 indices, revealing universal behavior and polynomial annealing time scaling of the VNA algorithm on portfolio optimization problems.", "date": "2025-07-09", "categories": []} {"id": "2405.09005", "title": "Cons-training Tensor Networks: Embedding and Optimization Over Discrete Linear Constraints", "abstract": "In this study, we introduce a novel family of tensor networks, termed constrained matrix product states (MPS), designed to incorporate exactly arbitrary discrete linear constraints, including inequalities, into sparse block structures. These tensor networks are particularly tailored for modeling distributions with support strictly over the feasible space, offering benefits such as reducing the search space in optimization problems, alleviating overfitting, improving training efficiency, and decreasing model size. Central to our approach is the concept of a quantum region, an extension of quantum numbers traditionally used in U(1) symmetric tensor networks, adapted to capture any linear constraint, including the unconstrained scenario. We further develop a novel canonical form for these new MPS, which allow for the merging and factorization of tensor blocks according to quantum region fusion rules and permit optimal truncation schemes. Utilizing this canonical form, we apply an unsupervised training strategy to optimize arbitrary objective functions subject to discrete linear constraints. Our method's efficacy is demonstrated by solving the quadratic knapsack problem, achieving superior performance compared to a leading nonlinear integer programming solver. Additionally, we analyze the complexity and scalability of our approach, demonstrating its potential in addressing complex constrained combinatorial optimization problems.", "date": "2024-05-15", "categories": []} {"id": "2406.18752", "title": "Near-Optimal Consistency-Robustness Trade-Offs for Learning-Augmented Online Knapsack Problems", "abstract": "This paper introduces a family of learning-augmented algorithms for online knapsack problems that achieve near Pareto-optimal consistency-robustness trade-offs through a simple combination of trusted learning-augmented and worst-case algorithms. Our approach relies on succinct, practical predictions -- single values or intervals estimating the minimum value of any item in an offline solution. Additionally, we propose a novel fractional-to-integral conversion procedure, offering new insights for online algorithm design.", "date": "2024-06-26", "categories": []} {"id": "2501.14568", "title": "Hybrid Quantum-Classical Multi-Agent Pathfinding", "abstract": "Multi-Agent Path Finding (MAPF) focuses on determining conflict-free paths for multiple agents navigating through a shared space to reach specified goal locations. This problem becomes computationally challenging, particularly when handling large numbers of agents, as frequently encountered in practical applications like coordinating autonomous vehicles. Quantum Computing (QC) is a promising candidate in overcoming such limits. However, current quantum hardware is still in its infancy and thus limited in terms of computing power and error robustness. In this work, we present the first optimal hybrid quantum-classical MAPF algorithms which are based on branch-andcut-and-price. QC is integrated by iteratively solving QUBO problems, based on conflict graphs. Experiments on actual quantum hardware and results on benchmark data suggest that our approach dominates previous QUBO formulationsand state-of-the-art MAPF solvers.", "date": "2025-01-24", "categories": []} {"id": "2507.05322", "title": "Dataless Neural Networks for Resource-Constrained Project Scheduling", "abstract": "Dataless neural networks represent a paradigm shift in applying neural architectures to combinatorial optimization problems, eliminating the need for training datasets by encoding problem instances directly into network parameters. Despite the pioneering work of Alkhouri et al. (2022) demonstrating the viability of dataless approaches for the Maximum Independent Set problem, our comprehensive literature review reveals that no published work has extended these methods to the Resource-Constrained Project Scheduling Problem (RCPSP). This paper addresses this gap by presenting the first dataless neural network approach for RCPSP, providing a complete mathematical framework that transforms discrete scheduling constraints into differentiable objectives suitable for gradient-based optimization. Our approach leverages smooth relaxations and automatic differentiation to unlock GPU parallelization for project scheduling, traditionally a domain of sequential algorithms. We detail the mathematical formulation for both precedence and renewable resource constraints, including a memory-efficient dense time-grid representation. Implementation and comprehensive experiments on PSPLIB benchmark instances (J30, J60, and J120) are currently underway, with empirical results to be reported in an updated version of this paper.", "date": "2025-07-07", "categories": []} {"id": "2507.06222", "title": "Deep Learning Optimization of Two-State Pinching Antennas Systems", "abstract": "The evolution of wireless communication systems requires flexible, energy-efficient, and cost-effective antenna technologies. Pinching antennas (PAs), which can dynamically control electromagnetic wave propagation through binary activation states, have recently emerged as a promising candidate. In this work, we investigate the problem of optimally selecting a subset of fixed-position PAs to activate in a waveguide, when the aim is to maximize the communication rate at a user terminal. Due to the complex interplay between antenna activation, waveguide-induced phase shifts, and power division, this problem is formulated as a combinatorial fractional 0-1 quadratic program. To efficiently solve this challenging problem, we use neural network architectures of varying complexity to learn activation policies directly from data, leveraging spatial features and signal structure. Furthermore, we incorporate user location uncertainty into our training and evaluation pipeline to simulate realistic deployment conditions. Simulation results demonstrate the effectiveness and robustness of the proposed models.", "date": "2025-07-08", "categories": []} {"id": "2507.03300", "title": "LRM-1B: Towards Large Routing Model", "abstract": "Vehicle routing problems (VRPs) are central to combinatorial optimization with significant practical implications. Recent advancements in neural combinatorial optimization (NCO) have demonstrated promising results by leveraging neural networks to solve VRPs, yet the exploration of model scaling within this domain remains underexplored. Inspired by the success of model scaling in large language models (LLMs), this study introduces a Large Routing Model with 1 billion parameters (LRM-1B), designed to address diverse VRP scenarios. We present a comprehensive evaluation of LRM-1B across multiple problem variants, distributions, and sizes, establishing state-of-the-art results. Our findings reveal that LRM-1B not only adapts to different VRP challenges but also showcases superior performance, outperforming existing models. Additionally, we explore the scaling behavior of neural routing models from 1M to 1B parameters. Our analysis confirms power-law between multiple model factors and performance, offering critical insights into the optimal configurations for foundation neural routing solvers.", "date": "2025-07-04", "categories": []} {"id": "2507.03775", "title": "Optimizing UAV Trajectories via a Simplified Close Enough TSP Approach", "abstract": "This article explores an approach to addressing the Close Enough Traveling Salesman Problem (CETSP). The objective is to streamline the mathematical formulation by introducing reformulations that approximate the Euclidean distances and simplify the objective function. Additionally, the use of convex sets in the constraint design offers computational benefits. The proposed methodology is empirically validated on real-world CETSP instances, with the aid of computational strategies such as a fragmented CPLEX-based approach. Results demonstrate its effectiveness in managing computational resources without compromising solution quality. Furthermore, the article analyzes the behavior of the proposed mathematical formulations, providing comprehensive insights into their performance.", "date": "2025-07-04", "categories": []} {"id": "2404.02476", "title": "Deep Reinforcement Learning for Traveling Purchaser Problems", "abstract": "The traveling purchaser problem (TPP) is an important combinatorial optimization problem with broad applications. Due to the coupling between routing and purchasing, existing works on TPPs commonly address route construction and purchase planning simultaneously, which, however, leads to exact methods with high computational cost and heuristics with sophisticated design but limited performance. In sharp contrast, we propose a novel approach based on deep reinforcement learning (DRL), which addresses route construction and purchase planning separately, while evaluating and optimizing the solution from a global perspective. The key components of our approach include a bipartite graph representation for TPPs to capture the market-product relations, and a policy network that extracts information from the bipartite graph and uses it to sequentially construct the route. One significant advantage of our framework is that we can efficiently construct the route using the policy network, and once the route is determined, the associated purchasing plan can be easily derived through linear programming, while, by leveraging DRL, we can train the policy network towards optimizing the global solution objective. Furthermore, by introducing a meta-learning strategy, the policy network can be trained stably on large-sized TPP instances, and generalize well across instances of varying sizes and distributions, even to much larger instances that are never seen during training. Experiments on various synthetic TPP instances and the TPPLIB benchmark demonstrate that our DRL-based approach can significantly outperform well-established TPP heuristics, reducing the optimality gap by 40%-90%, and also showing an advantage in runtime, especially on large-sized instances.", "date": "2024-04-03", "categories": []} {"id": "2503.00033", "title": "optimizn: a Python Library for Developing Customized Optimization Algorithms", "abstract": "Combinatorial optimization problems are prevalent across a wide variety of domains. These problems are often nuanced, their optimal solutions might not be efficiently obtainable, and they may require lots of time and compute resources to solve (they are NP-hard). It follows that the best course of action for solving these problems is to use general optimization algorithm paradigms to quickly and easily develop algorithms that are customized to these problems and can produce good solutions in a reasonable amount of time. In this paper, we present optimizn, a Python library for developing customized optimization algorithms under general optimization algorithm paradigms (simulated annealing, branch and bound). Additionally, optimizn offers continuous training, with which users can run their algorithms on a regular cadence, retain the salient aspects of previous runs, and use them in subsequent runs to potentially produce solutions that get closer and closer to optimality. An earlier version of this paper was peer reviewed and published internally at Microsoft.", "date": "2025-02-25", "categories": []} {"id": "2507.01462", "title": "Quantum-Assisted Automatic Path-Planning for Robotic Quality Inspection in Industry 4.0", "abstract": "This work explores the application of hybrid quantum-classical algorithms to optimize robotic inspection trajectories derived from Computer-Aided Design (CAD) models in industrial settings. By modeling the task as a 3D variant of the Traveling Salesman Problem, incorporating incomplete graphs and open-route constraints, this study evaluates the performance of two D-Wave-based solvers against classical methods such as GUROBI and Google OR-Tools. Results across five real-world cases demonstrate competitive solution quality with significantly reduced computation times, highlighting the potential of quantum approaches in automation under Industry 4.0.", "date": "2025-07-02", "categories": []} {"id": "2507.01638", "title": "Customized Exploration of Landscape Features Driving Multi-Objective Combinatorial Optimization Performance", "abstract": "We present an analysis of landscape features for predicting the performance of multi-objective combinatorial optimization algorithms. We consider features from the recently proposed compressed Pareto Local Optimal Solutions Networks (C-PLOS-net) model of combinatorial landscapes. The benchmark instances are a set of rmnk-landscapes with 2 and 3 objectives and various levels of ruggedness and objective correlation. We consider the performance of three algorithms -- Pareto Local Search (PLS), Global Simple EMO Optimizer (GSEMO), and Non-dominated Sorting Genetic Algorithm (NSGA-II) - using the resolution and hypervolume metrics. Our tailored analysis reveals feature combinations that influence algorithm performance specific to certain landscapes. This study provides deeper insights into feature importance, tailored to specific rmnk-landscapes and algorithms.", "date": "2025-07-02", "categories": []} {"id": "2408.00790", "title": "Improving Air Mobility for Pre-Disaster Planning with Neural Network Accelerated Genetic Algorithm", "abstract": "Weather disaster related emergency operations pose a great challenge to air mobility in both aircraft and airport operations, especially when the impact is gradually approaching. We propose an optimized framework for adjusting airport operational schedules for such pre-disaster scenarios. We first, aggregate operational data from multiple airports and then determine the optimal count of evacuation flights to maximize the impacted airport's outgoing capacity without impeding regular air traffic. We then propose a novel Neural Network (NN) accelerated Genetic Algorithm(GA) for evacuation planning. Our experiments show that integration yielded comparable results but with smaller computational overhead. We find that the utilization of a NN enhances the efficiency of a GA, facilitating more rapid convergence even when operating with a reduced population size. This effectiveness persists even when the model is trained on data from airports different from those under test.", "date": "2024-07-17", "categories": []} {"id": "2506.23793", "title": "Advancing Learnable Multi-Agent Pathfinding Solvers with Active Fine-Tuning", "abstract": "Multi-agent pathfinding (MAPF) is a common abstraction of multi-robot trajectory planning problems, where multiple homogeneous robots simultaneously move in the shared environment. While solving MAPF optimally has been proven to be NP-hard, scalable, and efficient, solvers are vital for real-world applications like logistics, search-and-rescue, etc. To this end, decentralized suboptimal MAPF solvers that leverage machine learning have come on stage. Building on the success of the recently introduced MAPF-GPT, a pure imitation learning solver, we introduce MAPF-GPT-DDG. This novel approach effectively fine-tunes the pre-trained MAPF model using centralized expert data. Leveraging a novel delta-data generation mechanism, MAPF-GPT-DDG accelerates training while significantly improving performance at test time. Our experiments demonstrate that MAPF-GPT-DDG surpasses all existing learning-based MAPF solvers, including the original MAPF-GPT, regarding solution quality across many testing scenarios. Remarkably, it can work with MAPF instances involving up to 1 million agents in a single environment, setting a new milestone for scalability in MAPF domains.", "date": "2025-06-30", "categories": []} {"id": "2411.01707", "title": "Large-Scale Multirobot Coverage Path Planning on Grids With Path Deconfliction", "abstract": "We study Multi-Robot Coverage Path Planning (MCPP) on a 4-neighbor 2D grid G, which aims to compute paths for multiple robots to cover all cells of G. Traditional approaches are limited as they first compute coverage trees on a quadrant coarsened grid H and then employ the Spanning Tree Coverage (STC) paradigm to generate paths on G, making them inapplicable to grids with partially obstructed 2x2 blocks. To address this limitation, we reformulate the problem directly on G, revolutionizing grid-based MCPP solving and establishing new NP-hardness results. We introduce Extended-STC (ESTC), a novel paradigm that extends STC to ensure complete coverage with bounded suboptimality, even when H includes partially obstructed blocks. Furthermore, we present LS-MCPP, a new algorithmic framework that integrates ESTC with three novel types of neighborhood operators within a local search strategy to optimize coverage paths directly on G. Unlike prior grid-based MCPP work, our approach also incorporates a versatile post-processing procedure that applies Multi-Agent Path Finding (MAPF) techniques to MCPP for the first time, enabling a fusion of these two important fields in multi-robot coordination. This procedure effectively resolves inter-robot conflicts and accommodates turning costs by solving a MAPF variant, making our MCPP solutions more practical for real-world applications. Extensive experiments demonstrate that our approach significantly improves solution quality and efficiency, managing up to 100 robots on grids as large as 256x256 within minutes of runtime. Validation with physical robots confirms the feasibility of our solutions under real-world conditions.", "date": "2024-11-03", "categories": []} {"id": "2501.18184", "title": "Genetic Algorithm with Innovative Chromosome Patterns in the Breeding Process", "abstract": "This paper proposes Genetic Algorithm with Border Trades (GAB), a novel modification of the standard genetic algorithm that enhances exploration by incorporating new chromosome patterns in the breeding process. This approach significantly mitigates premature convergence and improves search diversity. Empirically, GAB achieves up to 8x higher fitness and 10x faster convergence on complex job scheduling problems compared to standard Genetic Algorithms, reaching average fitness scores of 888 versus 106 in under 20 seconds. On the classic Flip-Flop problem, GAB consistently finds optimal or near-optimal solutions in fewer generations, even as input sizes scale to thousands of bits. These results highlight GAB as a highly effective and computationally efficient alternative for solving large-scale combinatorial optimization problems.", "date": "2025-01-30", "categories": []} {"id": "2506.20401", "title": "Smart Ride and Delivery Services with Electric Vehicles: Leveraging Bidirectional Charging for Profit Optimisation", "abstract": "With the rising popularity of electric vehicles (EVs), modern service systems, such as ride-hailing delivery services, are increasingly integrating EVs into their operations. Unlike conventional vehicles, EVs often have a shorter driving range, necessitating careful consideration of charging when fulfilling requests. With recent advances in Vehicle-to-Grid (V2G) technology - allowing EVs to also discharge energy back to the grid - new opportunities and complexities emerge. We introduce the Electric Vehicle Orienteering Problem with V2G (EVOP-V2G): a profit-maximization problem where EV drivers must select customer requests or orders while managing when and where to charge or discharge. This involves navigating dynamic electricity prices, charging station selection, and route constraints. We formulate the problem as a Mixed Integer Programming (MIP) model and propose two near-optimal metaheuristic algorithms: one evolutionary (EA) and the other based on large neighborhood search (LNS). Experiments on real-world data show our methods can double driver profits compared to baselines, while maintaining near-optimal performance on small instances and excellent scalability on larger ones. Our work highlights a promising path toward smarter, more profitable EV-based mobility systems that actively support the energy grid.", "date": "2025-06-25", "categories": []} {"id": "2506.15196", "title": "HeurAgenix: Leveraging LLMs for Solving Complex Combinatorial Optimization Challenges", "abstract": "Heuristic algorithms play a vital role in solving combinatorial optimization (CO) problems, yet traditional designs depend heavily on manual expertise and struggle to generalize across diverse instances. We introduce \\textbf{HeurAgenix}, a two-stage hyper-heuristic framework powered by large language models (LLMs) that first evolves heuristics and then selects among them automatically. In the heuristic evolution phase, HeurAgenix leverages an LLM to compare seed heuristic solutions with higher-quality solutions and extract reusable evolution strategies. During problem solving, it dynamically picks the most promising heuristic for each problem state, guided by the LLM's perception ability. For flexibility, this selector can be either a state-of-the-art LLM or a fine-tuned lightweight model with lower inference cost. To mitigate the scarcity of reliable supervision caused by CO complexity, we fine-tune the lightweight heuristic selector with a dual-reward mechanism that jointly exploits singals from selection preferences and state perception, enabling robust selection under noisy annotations. Extensive experiments on canonical benchmarks show that HeurAgenix not only outperforms existing LLM-based hyper-heuristics but also matches or exceeds specialized solvers. Code is available at https://github.com/microsoft/HeurAgenix.", "date": "2025-06-18", "categories": []} {"id": "2506.19375", "title": "Path Learning with Trajectory Advantage Regression", "abstract": "In this paper, we propose trajectory advantage regression, a method of offline path learning and path attribution based on reinforcement learning. The proposed method can be used to solve path optimization problems while algorithmically only solving a regression problem.", "date": "2025-06-24", "categories": []} {"id": "2502.11422", "title": "Planning of Heuristics: Strategic Planning on Large Language Models with Monte Carlo Tree Search for Automating Heuristic Optimization", "abstract": "Heuristics have achieved great success in solving combinatorial optimization problems~(COPs). However, heuristics designed by humans require too much domain knowledge and testing time. Since Large Language Models~(LLMs) possess strong capabilities to understand and generate content with a knowledge base that covers various domains, they offer potential ways to automatically optimize heuristics. To this end, we propose Planning of Heuristics~(PoH), an optimization method that integrates LLM self-reflection with Monte Carlo Tree Search, a well-known planning algorithm. PoH iteratively refines generated heuristics by evaluating their performance and providing improvement suggestions. Our method enables to iteratively evaluate the generated heuristics~(states) and improve them based on the improvement suggestions~(actions) and evaluation results~(rewards), by effectively simulating future states to search for paths with higher rewards. In this paper, we apply PoH to solve the Traveling Salesman Problem and the Flow Shop Scheduling Problem. The experimental results show that PoH outperforms hand-crafted heuristics and other Automatic Heuristic Design methods based on LLMs, and achieves the state-of-the-art performance in automating heuristic optimization with LLMs to solve tested COPs, especially with large sizes.", "date": "2025-02-17", "categories": []} {"id": "2506.15774", "title": "Advancing Stochastic 3-SAT Solvers by Dissipating Oversatisfied Constraints", "abstract": "We introduce and benchmark a stochastic local search heuristic for the NP-complete satisfiability problem 3-SAT that drastically outperforms existing solvers in the notoriously difficult realm of critically hard instances. Our construction is based on the crucial observation that well established previous approaches such as WalkSAT are prone to get stuck in local minima that are distinguished from true solutions by a larger number of oversatisfied combinatorial constraints. To address this issue, the proposed algorithm, coined DOCSAT, dissipates oversatisfied constraints (DOC), i.e. reduces their unfavorable abundance so as to render them critical. We analyze and benchmark our algorithm on a randomly generated sample of hard but satisfiable 3-SAT instances with varying problem sizes up to N=15000. Quite remarkably, we find that DOCSAT outperforms both WalkSAT and other well known algorithms including the complete solver Kissat, even when comparing its ability to solve the hardest quintile of the sample to the average performance of its competitors. The essence of DOCSAT may be seen as a way of harnessing statistical structure beyond the primary cost function of a combinatorial problem to avoid or escape local minima traps in stochastic local search, which opens avenues for generalization to other optimization problems.", "date": "2025-06-18", "categories": []} {"id": "2506.16428", "title": "EFormer: An Effective Edge-based Transformer for Vehicle Routing Problems", "abstract": "Recent neural heuristics for the Vehicle Routing Problem (VRP) primarily rely on node coordinates as input, which may be less effective in practical scenarios where real cost metrics-such as edge-based distances-are more relevant. To address this limitation, we introduce EFormer, an Edge-based Transformer model that uses edge as the sole input for VRPs. Our approach employs a precoder module with a mixed-score attention mechanism to convert edge information into temporary node embeddings. We also present a parallel encoding strategy characterized by a graph encoder and a node encoder, each responsible for processing graph and node embeddings in distinct feature spaces, respectively. This design yields a more comprehensive representation of the global relationships among edges. In the decoding phase, parallel context embedding and multi-query integration are used to compute separate attention mechanisms over the two encoded embeddings, facilitating efficient path construction. We train EFormer using reinforcement learning in an autoregressive manner. Extensive experiments on the Traveling Salesman Problem (TSP) and Capacitated Vehicle Routing Problem (CVRP) reveal that EFormer outperforms established baselines on synthetic datasets, including large-scale and diverse distributions. Moreover, EFormer demonstrates strong generalization on real-world instances from TSPLib and CVRPLib. These findings confirm the effectiveness of EFormer's core design in solving VRPs.", "date": "2025-06-19", "categories": []} {"id": "2506.16732", "title": "On Training-Test (Mis)alignment in Unsupervised Combinatorial Optimization: Observation, Empirical Exploration, and Analysis", "abstract": "In unsupervised combinatorial optimization (UCO), during training, one aims to have continuous decisions that are promising in a probabilistic sense for each training instance, which enables end-to-end training on initially discrete and non-differentiable problems. At the test time, for each test instance, starting from continuous decisions, derandomization is typically applied to obtain the final deterministic decisions. Researchers have developed more and more powerful test-time derandomization schemes to enhance the empirical performance and the theoretical guarantee of UCO methods. However, we notice a misalignment between training and testing in the existing UCO methods. Consequently, lower training losses do not necessarily entail better post-derandomization performance, even for the training instances without any data distribution shift. Empirically, we indeed observe such undesirable cases. We explore a preliminary idea to better align training and testing in UCO by including a differentiable version of derandomization into training. Our empirical exploration shows that such an idea indeed improves training-test alignment, but also introduces nontrivial challenges into training.", "date": "2025-06-20", "categories": []} {"id": "2506.16924", "title": "Real-Time Black-Box Optimization for Dynamic Discrete Environments Using Embedded Ising Machines", "abstract": "Many real-time systems require the optimization of discrete variables. Black-box optimization (BBO) algorithms and multi-armed bandit (MAB) algorithms perform optimization by repeatedly taking actions and observing the corresponding instant rewards without any prior knowledge. Recently, a BBO method using an Ising machine has been proposed to find the best action that is represented by a combination of discrete values and maximizes the instant reward in static environments. In contrast, dynamic environments, where real-time systems operate, necessitate MAB algorithms that maximize the average reward over multiple trials. However, due to the enormous number of actions resulting from the combinatorial nature of discrete optimization, conventional MAB algorithms cannot effectively optimize dynamic, discrete environments. Here, we show a heuristic MAB method for dynamic, discrete environments by extending the BBO method, in which an Ising machine effectively explores the actions while considering interactions between variables and changes in dynamic environments. We demonstrate the dynamic adaptability of the proposed method in a wireless communication system with moving users.", "date": "2025-06-20", "categories": []} {"id": "2406.13023", "title": "$k$-Submodular Interdiction Problems under Distributional Risk-Receptiveness and Robustness: Application to Machine Learning", "abstract": "We study submodular optimization in adversarial context, applicable to machine learning problems such as feature selection using data susceptible to uncertainties and attacks. We focus on Stackelberg games between an attacker (or interdictor) and a defender where the attacker aims to minimize the defender's objective of maximizing a $k$-submodular function. We allow uncertainties arising from the success of attacks and inherent data noise, and address challenges due to incomplete knowledge of the probability distribution of random parameters. Specifically, we introduce Distributionally Robust $k$-Submodular Interdiction Problem (DRO $k$-SIP) and Distributionally Risk-Receptive $k$-Submodular Interdiction Problem (DRR $k$-SIP) along with finitely convergent exact algorithms for solving them. When solving the DRO $k$-SIP, the attacker optimizes their expected payoff with respect to the worst-case probability distribution within the ambiguity set, and thereby have robust attack strategies despite distributional ambiguity. In contrast, the DRR $k$-SIP identifies attacker strategies with the best-case probability distribution, and identifies critical vulnerabilities for the defender. The optimal values derived from both DRO $k$-SIP and DRR $k$-SIP offer a confidence interval-like range for the expected value of the defender's objective function, capturing distributional ambiguity. We conduct computational experiments on instances of feature selection and sensor placement problems, using Wisconsin breast cancer data and synthetic data, respectively.", "date": "2024-06-18", "categories": []} {"id": "2506.13781", "title": "Solving the Job Shop Scheduling Problem with Graph Neural Networks: A Customizable Reinforcement Learning Environment", "abstract": "The job shop scheduling problem is an NP-hard combinatorial optimization problem relevant to manufacturing and timetabling. Traditional approaches use priority dispatching rules based on simple heuristics. Recent work has attempted to replace these with deep learning models, particularly graph neural networks (GNNs), that learn to assign priorities from data. However, training such models requires customizing numerous factors: graph representation, node features, action space, and reward functions. The lack of modular libraries for experimentation makes this research time-consuming. This work introduces JobShopLib, a modular library that allows customizing these factors and creating new components with its reinforcement learning environment. We trained several dispatchers through imitation learning to demonstrate the environment's utility. One model outperformed various graph-based dispatchers using only individual operation features, highlighting the importance of feature customization. Our GNN model achieved near state-of-the-art results on large-scale problems. These results suggest significant room for improvement in developing such models. JobShopLib provides the necessary tools for future experimentation.", "date": "2025-06-10", "categories": []} {"id": "2506.13810", "title": "Bridging Pattern-Aware Complexity with NP-Hard Optimization: A Unifying Framework and Empirical Study", "abstract": "NP hard optimization problems like the Traveling Salesman Problem (TSP) defy efficient solutions in the worst case, yet real-world instances often exhibit exploitable patterns. We propose a novel patternaware complexity framework that quantifies and leverages structural regularities e.g., clustering, symmetry to reduce effective computational complexity across domains, including financial forecasting and LLM optimization. With rigorous definitions, theorems, and a meta learning driven solver pipeline, we introduce metrics like Pattern Utilization Efficiency (PUE) and achieve up to 79 percent solution quality gains in TSP benchmarks (22 to 2392 cities). Distinct from theoretical NP hardness, our approach offers a unified, practical lens for pattern-driven efficiency.", "date": "2025-03-12", "categories": []} {"id": "2505.13405", "title": "A Dataless Reinforcement Learning Approach to Rounding Hyperplane Optimization for Max-Cut", "abstract": "The Maximum Cut (MaxCut) problem is NP-Complete, and obtaining its optimal solution is NP-hard in the worst case. As a result, heuristic-based algorithms are commonly used, though their design often requires significant domain expertise. More recently, learning-based methods trained on large (un)labeled datasets have been proposed; however, these approaches often struggle with generalizability and scalability. A well-known approximation algorithm for MaxCut is the Goemans-Williamson (GW) algorithm, which relaxes the Quadratic Unconstrained Binary Optimization (QUBO) formulation into a semidefinite program (SDP). The GW algorithm then applies hyperplane rounding by uniformly sampling a random hyperplane to convert the SDP solution into binary node assignments. In this paper, we propose a training-data-free approach based on a non-episodic reinforcement learning formulation, in which an agent learns to select improved rounding hyperplanes that yield better cuts than those produced by the GW algorithm. By optimizing over a Markov Decision Process (MDP), our method consistently achieves better cuts across large-scale graphs with varying densities and degree distributions.", "date": "2025-05-19", "categories": []} {"id": "2506.12495", "title": "Automated Heuristic Design for Unit Commitment Using Large Language Models", "abstract": "The Unit Commitment (UC) problem is a classic challenge in the optimal scheduling of power systems. Years of research and practice have shown that formulating reasonable unit commitment plans can significantly improve the economic efficiency of power systems' operations. In recent years, with the introduction of technologies such as machine learning and the Lagrangian relaxation method, the solution methods for the UC problem have become increasingly diversified, but still face challenges in terms of accuracy and robustness. This paper proposes a Function Space Search (FunSearch) method based on large language models. This method combines pre-trained large language models and evaluators to creatively generate solutions through the program search and evolution process while ensuring their rationality. In simulation experiments, a case of unit commitment with \\(10\\) units is used mainly. Compared to the genetic algorithm, the results show that FunSearch performs better in terms of sampling time, evaluation time, and total operating cost of the system, demonstrating its great potential as an effective tool for solving the UC problem.", "date": "2025-06-14", "categories": []} {"id": "2506.06121", "title": "Decomposability-Guaranteed Cooperative Coevolution for Large-Scale Itinerary Planning", "abstract": "Large-scale itinerary planning is a variant of the traveling salesman problem, aiming to determine an optimal path that maximizes the collected points of interest (POIs) scores while minimizing travel time and cost, subject to travel duration constraints. This paper analyzes the decomposability of large-scale itinerary planning, proving that strict decomposability is difficult to satisfy, and introduces a weak decomposability definition based on a necessary condition, deriving the corresponding graph structures that fulfill this property. With decomposability guaranteed, we propose a novel multi-objective cooperative coevolutionary algorithm for large-scale itinerary planning, addressing the challenges of component imbalance and interactions. Specifically, we design a dynamic decomposition strategy based on the normalized fitness within each component, define optimization potential considering component scale and contribution, and develop a computational resource allocation strategy. Finally, we evaluate the proposed algorithm on a set of real-world datasets. Comparative experiments with state-of-the-art multi-objective itinerary planning algorithms demonstrate the superiority of our approach, with performance advantages increasing as the problem scale grows.", "date": "2025-06-06", "categories": []} {"id": "2506.11052", "title": "ACCORD: Autoregressive Constraint-satisfying Generation for COmbinatorial Optimization with Routing and Dynamic attention", "abstract": "Large Language Models (LLMs) have demonstrated impressive reasoning capabilities, yet their direct application to NP-hard combinatorial problems (CPs) remains underexplored. In this work, we systematically investigate the reasoning abilities of LLMs on a variety of NP-hard combinatorial optimization tasks and introduce ACCORD: Autoregressive Constraint-satisfying generation for COmbinatorial optimization with Routing and Dynamic attention. ACCORD features a novel dataset representation and model architecture that leverage the autoregressive nature of LLMs to dynamically enforce feasibility constraints, coupled with attention-based routing to activate problem-specific LoRA modules. We also present the ACCORD-90k supervised dataset, covering six NP-hard combinatorial problems: TSP, VRP, Knapsack, FlowShop, JSSP, and BinPacking. Extensive experiments demonstrate that our ACCORD model, built on an 8B-parameter Llama backbone, consistently outperforms standard prompting and input-output methods, even when compared to much larger LLMs, such as gpt-4. Ablation studies further show that our output structure enhances solution feasibility. To the best of our knowledge, this is the first large-scale, end-to-end framework for exploring the applications of LLMs to a broad spectrum of combinatorial optimization problems. The codes are publicly available at https://github.com/starjob42/ACCORD", "date": "2025-05-22", "categories": []} {"id": "2506.11057", "title": "STRCMP: Integrating Graph Structural Priors with Language Models for Combinatorial Optimization", "abstract": "Combinatorial optimization (CO) problems, central to operation research and theoretical computer science, present significant computational challenges due to their NP-hard nature. While large language models (LLMs) have emerged as promising tools for CO--either by directly generating solutions or synthesizing solver-specific codes--existing approaches often neglect critical structural priors inherent to CO problems, leading to suboptimality and iterative inefficiency. Inspired by human experts' success in leveraging CO structures for algorithm design, we propose STRCMP, a novel structure-aware LLM-based algorithm discovery framework that systematically integrates structure priors to enhance solution quality and solving efficiency. Our framework combines a graph neural network (GNN) for extracting structural embeddings from CO instances with an LLM conditioned on these embeddings to identify high-performing algorithms in the form of solver-specific codes. This composite architecture ensures syntactic correctness, preserves problem topology, and aligns with natural language objectives, while an evolutionary refinement process iteratively optimizes generated algorithm. Extensive evaluations across Mixed Integer Linear Programming and Boolean Satisfiability problems, using nine benchmark datasets, demonstrate that our proposed STRCMP outperforms five strong neural and LLM-based methods by a large margin, in terms of both solution optimality and computational efficiency. The code and learned model will be publicly available upon the acceptance of the paper.", "date": "2025-05-22", "categories": []} {"id": "2401.12533", "title": "Near-Optimal Algorithms for Constrained k-Center Clustering with Instance-level Background Knowledge", "abstract": "Center-based clustering has attracted significant research interest from both theory and practice. In many practical applications, input data often contain background knowledge that can be used to improve clustering results. In this work, we build on widely adopted $k$-center clustering and model its input background knowledge as must-link (ML) and cannot-link (CL) constraint sets. However, most clustering problems including $k$-center are inherently $\\mathcal{NP}$-hard, while the more complex constrained variants are known to suffer severer approximation and computation barriers that significantly limit their applicability. By employing a suite of techniques including reverse dominating sets, linear programming (LP) integral polyhedron, and LP duality, we arrive at the first efficient approximation algorithm for constrained $k$-center with the best possible ratio of 2. We also construct competitive baseline algorithms and empirically evaluate our approximation algorithm against them on a variety of real datasets. The results validate our theoretical findings and demonstrate the great advantages of our algorithm in terms of clustering cost, clustering quality, and running time.", "date": "2024-01-23", "categories": []} {"id": "2412.14409", "title": "Multi-task Representation Learning for Mixed Integer Linear Programming", "abstract": "Mixed Integer Linear Programs (MILPs) are highly flexible and powerful tools for modeling and solving complex real-world combinatorial optimization problems. Recently, machine learning (ML)-guided approaches have demonstrated significant potential in improving MILP-solving efficiency. However, these methods typically rely on separate offline data collection and training processes, which limits their scalability and adaptability. This paper introduces the first multi-task learning framework for ML-guided MILP solving. The proposed framework provides MILP embeddings helpful in guiding MILP solving across solvers (e.g., Gurobi and SCIP) and across tasks (e.g., Branching and Solver configuration). Through extensive experiments on three widely used MILP benchmarks, we demonstrate that our multi-task learning model performs similarly to specialized models within the same distribution. Moreover, it significantly outperforms them in generalization across problem sizes and tasks.", "date": "2024-12-18", "categories": []} {"id": "2502.10353", "title": "Assortment Optimization for Patient-Provider Matching", "abstract": "Rising provider turnover results in frequently needing to rematch patients with available providers. However, the rematching process is cumbersome for both patients and health systems, resulting in labor-intensive and ad hoc reassignments. We propose a novel patient-provider matching approach to address this issue by offering patients limited provider menus. The goal is to maximize match quality across the system while preserving patient choice. We frame this as a novel variant of assortment optimization, where patient-specific provider menus are offered upfront, and patients respond in a random sequence to make their selections. This hybrid offline-online setting is understudied in previous literature and captures system dynamics across various domains. We first demonstrate that a greedy baseline policy--which offers all providers to all patients--can maximize the match rate but lead to low-quality matches. Based on this, we construct a set of policies and demonstrate that the best policy depends on problem specifics, such as a patient's willingness to match and the ratio of patients to providers. On real-world data, our proposed policy improves average match quality by 13% over a greedy solution by tailoring assortments based on patient characteristics. Our analysis reveals a tradeoff between menu size and system-wide match quality, highlighting the value of balancing patient choice with centralized planning.", "date": "2025-02-14", "categories": []} {"id": "2502.11607", "title": "GraphThought: Graph Combinatorial Optimization with Thought Generation", "abstract": "Graph combinatorial optimization (GCO) problems are central to domains like logistics and bioinformatics. While traditional solvers dominate, large language models (LLMs) offer new possibilities for structured reasoning, yet struggle with complex GCO tasks requiring rigorous combinatorial analysis and multi-step deduction, often producing hallucinated steps. We first formalize the Optimal Thoughts Design (OTD) problem, which provides a structured guidance for producing high-quality intermediate reasoning steps. Building on this formulation, we introduce GraphThought, a novel framework that generates effective reasoning sequences through either heuristic-guided forward search or solver-aligned backward reasoning. By fine-tuning LLMs on these structured thought sequences, we develop Llama-GT, an 8B-parameter model that achieves state-of-the-art performance on the GraphArena benchmark, outperforming significantly larger models like DeepSeek-V3. Our results demonstrate that when scaffolded with structured reasoning priors, principled thought generation can significantly enhance LLM performance on GCO tasks without requiring increased model scale.", "date": "2025-02-17", "categories": []} {"id": "2505.13522", "title": "A Heuristic Algorithm Based on Beam Search and Iterated Local Search for the Maritime Inventory Routing Problem", "abstract": "Maritime Inventory Routing Problem (MIRP) plays a crucial role in the integration of global maritime commerce levels. However, there are still no well-established methodologies capable of efficiently solving large MIRP instances or their variants due to the high complexity of the problem. The adoption of exact methods, typically based on Mixed Integer Programming (MIP), for daily operations is nearly impractical due to the CPU time required, as planning must be executed multiple times while ensuring high-quality results within acceptable time limits. Non-MIP-based heuristics are less frequently applied due to the highly constrained nature of the problem, which makes even the construction of an effective initial solution challenging. Papageorgiou et al. (2014) introduced a single-product MIRP as the foundation for MIRPLib, aiming to provide a collection of publicly available benchmark instances. However, only a few studies that propose new methodologies have been published since then. To encourage the use of MIRPLib and facilitate result comparisons, this study presents a heuristic approach that does not rely on mathematical optimization techniques to solve a deterministic, finite-horizon, single-product MIRP. The proposed heuristic combines a variation of a Beam Search algorithm with an Iterated Local Search procedure. Among the 72 instances tested, the developed methodology can improve the best-known solution for 19 instances within an acceptable CPU time.", "date": "2025-05-17", "categories": []} {"id": "2405.20525", "title": "Comparing Quantum Annealing and Spiking Neuromorphic Computing for Sampling Binary Sparse Coding QUBO Problems", "abstract": "We consider the problem of computing a sparse binary representation of an image. To be precise, given an image and an overcomplete, non-orthonormal basis, we aim to find a sparse binary vector indicating the minimal set of basis vectors that when added together best reconstruct the given input. We formulate this problem with an $L_2$ loss on the reconstruction error, and an $L_0$ (or, equivalently, an $L_1$) loss on the binary vector enforcing sparsity. This yields a quadratic unconstrained binary optimization problem (QUBO), whose optimal solution(s) in general is NP-hard to find. The contribution of this work is twofold. First, we solve the sparse representation QUBOs by solving them both on a D-Wave quantum annealer with Pegasus chip connectivity via minor embedding, as well as on the Intel Loihi 2 spiking neuromorphic processor using a stochastic Non-equilibrium Boltzmann Machine (NEBM). Second, we deploy Quantum Evolution Monte Carlo with Reverse Annealing and iterated warm starting on Loihi 2 to evolve the solution quality from the respective machines. The solutions are benchmarked against simulated annealing, a classical heuristic, and the optimal solutions are computed using CPLEX. Iterated reverse quantum annealing performs similarly to simulated annealing, although simulated annealing is always able to sample the optimal solution whereas quantum annealing was not always able to. The Loihi 2 solutions that are sampled are on average more sparse than the solutions from any of the other methods. We demonstrate that both quantum annealing and neuromorphic computing are suitable for binary sparse coding QUBOs, and that Loihi 2 outperforms a D-Wave quantum annealer standard linear-schedule anneal, while iterated reverse quantum annealing performs much better than both unmodified linear-schedule quantum annealing and iterated warm starting on Loihi 2.", "date": "2024-05-30", "categories": []} {"id": "2506.08424", "title": "SHIELD: Multi-task Multi-distribution Vehicle Routing Solver with Sparsity and Hierarchy", "abstract": "Recent advances toward foundation models for routing problems have shown great potential of a unified deep model for various VRP variants. However, they overlook the complex real-world customer distributions. In this work, we advance the Multi-Task VRP (MTVRP) setting to the more realistic yet challenging Multi-Task Multi-Distribution VRP (MTMDVRP) setting, and introduce SHIELD, a novel model that leverages both sparsity and hierarchy principles. Building on a deeper decoder architecture, we first incorporate the Mixture-of-Depths (MoD) technique to enforce sparsity. This improves both efficiency and generalization by allowing the model to dynamically select nodes to use or skip each decoder layer, providing the needed capacity to adaptively allocate computation for learning the task/distribution specific and shared representations. We also develop a context-based clustering layer that exploits the presence of hierarchical structures in the problems to produce better local representations. These two designs inductively bias the network to identify key features that are common across tasks and distributions, leading to significantly improved generalization on unseen ones. Our empirical results demonstrate the superiority of our approach over existing methods on 9 real-world maps with 16 VRP variants each.", "date": "2025-06-10", "categories": []} {"id": "2506.09404", "title": "Synergizing Reinforcement Learning and Genetic Algorithms for Neural Combinatorial Optimization", "abstract": "Combinatorial optimization problems are notoriously challenging due to their discrete structure and exponentially large solution space. Recent advances in deep reinforcement learning (DRL) have enabled the learning heuristics directly from data. However, DRL methods often suffer from limited exploration and susceptibility to local optima. On the other hand, evolutionary algorithms such as Genetic Algorithms (GAs) exhibit strong global exploration capabilities but are typically sample inefficient and computationally intensive. In this work, we propose the Evolutionary Augmentation Mechanism (EAM), a general and plug-and-play framework that synergizes the learning efficiency of DRL with the global search power of GAs. EAM operates by generating solutions from a learned policy and refining them through domain-specific genetic operations such as crossover and mutation. These evolved solutions are then selectively reinjected into the policy training loop, thereby enhancing exploration and accelerating convergence. We further provide a theoretical analysis that establishes an upper bound on the KL divergence between the evolved solution distribution and the policy distribution, ensuring stable and effective policy updates. EAM is model-agnostic and can be seamlessly integrated with state-of-the-art DRL solvers such as the Attention Model, POMO, and SymNCO. Extensive results on benchmark problems (e.g., TSP, CVRP, PCTSP, and OP) demonstrate that EAM significantly improves both solution quality and training efficiency over competitive baselines.", "date": "2025-06-11", "categories": []} {"id": "2502.14760", "title": "EquivaMap: Leveraging LLMs for Automatic Equivalence Checking of Optimization Formulations", "abstract": "A fundamental problem in combinatorial optimization is identifying equivalent formulations. Despite the growing need for automated equivalence checks -- driven, for example, by optimization copilots, which generate problem formulations from natural language descriptions -- current approaches rely on simple heuristics that fail to reliably check formulation equivalence. Inspired by Karp reductions, in this work we introduce Quasi-Karp equivalence, a formal criterion for determining when two optimization formulations are equivalent based on the existence of a mapping between their decision variables. We propose EquivaMap, a framework that leverages large language models to automatically discover such mappings for scalable, reliable equivalence checking, with a verification stage that ensures mapped solutions preserve feasibility and optimality without additional solver calls. To evaluate our approach, we construct EquivaFormulation, the first open-source dataset of equivalent optimization formulations, generated by applying transformations such as adding slack variables or valid inequalities to existing formulations. Empirically, EquivaMap significantly outperforms existing methods, achieving substantial improvements in correctly identifying formulation equivalence.", "date": "2025-02-20", "categories": []} {"id": "2406.19532", "title": "Differentiable Quadratic Optimization For The Maximum Independent Set Problem", "abstract": "Combinatorial Optimization (CO) addresses many important problems, including the challenging Maximum Independent Set (MIS) problem. Alongside exact and heuristic solvers, differentiable approaches have emerged, often using continuous relaxations of ReLU-based or quadratic objectives. Noting that an MIS in a graph is a Maximum Clique (MC) in its complement, we propose a new quadratic formulation for MIS by incorporating an MC term, improving convergence and exploration. We show that every maximal independent set corresponds to a local minimizer, derive conditions with respect to the MIS size, and characterize stationary points. To tackle the non-convexity of the objective, we propose optimizing several initializations in parallel using momentum-based gradient descent, complemented by an efficient MIS checking criterion derived from our theory. We dub our method as parallelized Clique-Informed Quadratic Optimization for MIS (pCQO-MIS). Our experimental results demonstrate the effectiveness of the proposed method compared to exact, heuristic, sampling, and data-centric approaches. Notably, our method avoids the out-of-distribution tuning and reliance on (un)labeled data required by data-centric methods, while achieving superior MIS sizes and competitive runtime relative to their inference time. Additionally, a key advantage of pCQO-MIS is that, unlike exact and heuristic solvers, the runtime scales only with the number of nodes in the graph, not the number of edges. Our code is available at the GitHub repository: https://github.com/ledenmat/pCQO-mis-benchmark/tree/refactor.", "date": "2024-06-27", "categories": []} {"id": "2406.19705", "title": "DISCO: Efficient Diffusion Solver for Large-Scale Combinatorial Optimization Problems", "abstract": "Combinatorial Optimization (CO) problems are fundamentally important in numerous real-world applications across diverse industries, characterized by entailing enormous solution space and demanding time-sensitive response. Despite recent advancements in neural solvers, their limited expressiveness struggles to capture the multi-modal nature of CO landscapes. While some research has shifted towards diffusion models, these models still sample solutions indiscriminately from the entire NP-complete solution space with time-consuming denoising processes, which limit their practicality for large problem scales. We propose DISCO, an efficient DIffusion Solver for large-scale Combinatorial Optimization problems that excels in both solution quality and inference speed. DISCO's efficacy is twofold: First, it enhances solution quality by constraining the sampling space to a more meaningful domain guided by solution residues, while preserving the multi-modal properties of the output distributions. Second, it accelerates the denoising process through an analytically solvable approach, enabling solution sampling with minimal reverse-time steps and significantly reducing inference time. DISCO delivers strong performance on large-scale Traveling Salesman Problems and challenging Maximal Independent Set benchmarks, with inference time up to 5.28 times faster than other diffusion alternatives. By incorporating a divide-and-conquer strategy, DISCO can well generalize to solve unseen-scale problem instances, even surpassing models specifically trained for those scales.", "date": "2024-06-28", "categories": []} {"id": "2506.06291", "title": "Improvement of Optimization using Learning Based Models in Mixed Integer Linear Programming Tasks", "abstract": "Mixed Integer Linear Programs (MILPs) are essential tools for solving planning and scheduling problems across critical industries such as construction, manufacturing, and logistics. However, their widespread adoption is limited by long computational times, especially in large-scale, real-time scenarios. To address this, we present a learning-based framework that leverages Behavior Cloning (BC) and Reinforcement Learning (RL) to train Graph Neural Networks (GNNs), producing high-quality initial solutions for warm-starting MILP solvers in Multi-Agent Task Allocation and Scheduling Problems. Experimental results demonstrate that our method reduces optimization time and variance compared to traditional techniques while maintaining solution quality and feasibility.", "date": "2025-05-17", "categories": []} {"id": "2506.06634", "title": "GELD: A Unified Neural Model for Efficiently Solving Traveling Salesman Problems Across Different Scales", "abstract": "The Traveling Salesman Problem (TSP) is a well-known combinatorial optimization problem with broad real-world applications. Recent advancements in neural network-based TSP solvers have shown promising results. Nonetheless, these models often struggle to efficiently solve both small- and large-scale TSPs using the same set of pre-trained model parameters, limiting their practical utility. To address this issue, we introduce a novel neural TSP solver named GELD, built upon our proposed broad global assessment and refined local selection framework. Specifically, GELD integrates a lightweight Global-view Encoder (GE) with a heavyweight Local-view Decoder (LD) to enrich embedding representation while accelerating the decision-making process. Moreover, GE incorporates a novel low-complexity attention mechanism, allowing GELD to achieve low inference latency and scalability to larger-scale TSPs. Additionally, we propose a two-stage training strategy that utilizes training instances of different sizes to bolster GELD's generalization ability. Extensive experiments conducted on both synthetic and real-world datasets demonstrate that GELD outperforms seven state-of-the-art models considering both solution quality and inference speed. Furthermore, GELD can be employed as a post-processing method to significantly elevate the quality of the solutions derived by existing neural TSP solvers via spending affordable additional computing time. Notably, GELD is shown as capable of solving TSPs with up to 744,710 nodes, first-of-its-kind to solve this large size TSP without relying on divide-and-conquer strategies to the best of our knowledge.", "date": "2025-06-07", "categories": []} {"id": "2506.07373", "title": "HyColor: An Efficient Heuristic Algorithm for Graph Coloring", "abstract": "The graph coloring problem (GCP) is a classic combinatorial optimization problem that aims to find the minimum number of colors assigned to vertices of a graph such that no two adjacent vertices receive the same color. GCP has been extensively studied by researchers from various fields, including mathematics, computer science, and biological science. Due to the NP-hard nature, many heuristic algorithms have been proposed to solve GCP. However, existing GCP algorithms focus on either small hard graphs or large-scale sparse graphs (with up to 10^7 vertices). This paper presents an efficient hybrid heuristic algorithm for GCP, named HyColor, which excels in handling large-scale sparse graphs while achieving impressive results on small dense graphs. The efficiency of HyColor comes from the following three aspects: a local decision strategy to improve the lower bound on the chromatic number; a graph-reduction strategy to reduce the working graph; and a k-core and mixed degree-based greedy heuristic for efficiently coloring graphs. HyColor is evaluated against three state-of-the-art GCP algorithms across four benchmarks, comprising three large-scale sparse graph benchmarks and one small dense graph benchmark, totaling 209 instances. The results demonstrate that HyColor consistently outperforms existing heuristic algorithms in both solution accuracy and computational efficiency for the majority of instances. Notably, HyColor achieved the best solutions in 194 instances (over 93%), with 34 of these solutions significantly surpassing those of other algorithms. Furthermore, HyColor successfully determined the chromatic number and achieved optimal coloring in 128 instances.", "date": "2025-06-09", "categories": []} {"id": "2502.05221", "title": "Blackout DIFUSCO", "abstract": "This study explores the integration of Blackout Diffusion into the DIFUSCO framework for combinatorial optimization, specifically targeting the Traveling Salesman Problem (TSP). Inspired by the success of discrete-time diffusion models (D3PM) in maintaining structural integrity, we extend the paradigm to a continuous-time framework, leveraging the unique properties of Blackout Diffusion. Continuous-time modeling introduces smoother transitions and refined control, hypothesizing enhanced solution quality over traditional discrete methods. We propose three key improvements to enhance the diffusion process. First, we transition from a discrete-time-based model to a continuous-time framework, providing a more refined and flexible formulation. Second, we refine the observation time scheduling to ensure a smooth and linear transformation throughout the diffusion process, allowing for a more natural progression of states. Finally, building upon the second improvement, we further enhance the reverse process by introducing finer time slices in regions that are particularly challenging for the model, thereby improving accuracy and stability in the reconstruction phase. Although the experimental results did not exceed the baseline performance, they demonstrate the effectiveness of these methods in balancing simplicity and complexity, offering new insights into diffusion-based combinatorial optimization. This work represents the first application of Blackout Diffusion to combinatorial optimization, providing a foundation for further advancements in this domain. * The code is available for review at https://github.com/Giventicket/BlackoutDIFUSCO.", "date": "2025-02-05", "categories": []} {"id": "2505.15862", "title": "Bandit based Dynamic Candidate Edge Selection in Solving Traveling Salesman Problems", "abstract": "Algorithms designed for routing problems typically rely on high-quality candidate edges to guide their search, aiming to reduce the search space and enhance the search efficiency. However, many existing algorithms, like the classical Lin-Kernighan-Helsgaun (LKH) algorithm for the Traveling Salesman Problem (TSP), often use predetermined candidate edges that remain static throughout local searches. This rigidity could cause the algorithm to get trapped in local optima, limiting its potential to find better solutions. To address this issue, we propose expanding the candidate sets to include other promising edges, providing them an opportunity for selection. Specifically, we incorporate multi-armed bandit models to dynamically select the most suitable candidate edges in each iteration, enabling LKH to make smarter choices and lead to improved solutions. Extensive experiments on multiple TSP benchmarks show the excellent performance of our method. Moreover, we employ this bandit-based method to LKH-3, an extension of LKH tailored for solving various TSP variant problems, and our method also significantly enhances LKH-3's performance across typical TSP variants.", "date": "2025-05-21", "categories": []} {"id": "2506.03974", "title": "A Generic Branch-and-Bound Algorithm for $\\ell_0$-Penalized Problems with Supplementary Material", "abstract": "We present a generic Branch-and-Bound procedure designed to solve L0-penalized optimization problems. Existing approaches primarily focus on quadratic losses and construct relaxations using \"Big-M\" constraints and/or L2-norm penalties. In contrast, our method accommodates a broader class of loss functions and allows greater flexibility in relaxation design through a general penalty term, encompassing existing techniques as special cases. We establish theoretical results ensuring that all key quantities required for the Branch-and-Bound implementation admit closed-form expressions under the general blanket assumptions considered in our work. Leveraging this framework, we introduce El0ps, an open-source Python solver with a plug-and-play workflow that enables user-defined losses and penalties in L0-penalized problems. Through extensive numerical experiments, we demonstrate that El0ps achieves state-of-the-art performance on classical instances and extends computational feasibility to previously intractable ones.", "date": "2025-06-04", "categories": []} {"id": "2502.12012", "title": "Evolving Hard Maximum Cut Instances for Quantum Approximate Optimization Algorithms", "abstract": "Variational quantum algorithms, such as the Recursive Quantum Approximate Optimization Algorithm (RQAOA), have become increasingly popular, offering promising avenues for employing Noisy Intermediate-Scale Quantum devices to address challenging combinatorial optimization tasks like the maximum cut problem. In this study, we utilize an evolutionary algorithm equipped with a unique fitness function. This approach targets hard maximum cut instances within the latent space of a Graph Autoencoder, identifying those that pose significant challenges or are particularly tractable for RQAOA, in contrast to the classic Goemans and Williamson algorithm. Our findings not only delineate the distinct capabilities and limitations of each algorithm but also expand our understanding of RQAOA's operational limits. Furthermore, the diverse set of graphs we have generated serves as a crucial benchmarking asset, emphasizing the need for more advanced algorithms to tackle combinatorial optimization challenges. Additionally, our results pave the way for new avenues in graph generation research, offering exciting opportunities for future explorations.", "date": "2025-01-30", "categories": []} {"id": "2506.02594", "title": "EALG: Evolutionary Adversarial Generation of Language Model-Guided Generators for Combinatorial Optimization", "abstract": "Generating challenging instances is crucial for the evaluation and advancement of combinatorial optimization solvers. In this work, we introduce EALG (Evolutionary Adversarial Generation of Language Model-Guided Generators), a novel framework that automates the co-evolution of optimization problem instances and their corresponding heuristic solvers using large language models (LLMs). EALG leverages a mutation-based adversarial approach that dynamically evolves instance generation procedures to create increasingly difficult problems, while simultaneously synthesizing adaptive heuristic algorithms through interactions with LLMs guided by algorithmic structure. Unlike existing approaches that focus solely on static benchmark creation or manual solver design, EALG provides a seamless pipeline from instance generation to solver synthesis. Experimental results demonstrate that EALG generates significantly harder instances than current benchmarks, and its synthesized solvers generalize effectively across a broad spectrum of combinatorial tasks. This work explores a new paradigm for combinatorial optimization that integrates instance generation with solver design, resulting in state-of-the-art performance.", "date": "2025-06-03", "categories": []} {"id": "2506.02746", "title": "Solving the Pod Repositioning Problem with Deep Reinforced Adaptive Large Neighborhood Search", "abstract": "The Pod Repositioning Problem (PRP) in Robotic Mobile Fulfillment Systems (RMFS) involves selecting optimal storage locations for pods returning from pick stations. This work presents an improved solution method that integrates Adaptive Large Neighborhood Search (ALNS) with Deep Reinforcement Learning (DRL). A DRL agent dynamically selects destroy and repair operators and adjusts key parameters such as destruction degree and acceptance thresholds during the search. Specialized heuristics for both operators are designed to reflect PRP-specific characteristics, including pod usage frequency and movement costs. Computational results show that this DRL-guided ALNS outperforms traditional approaches such as cheapest-place, fixed-place, binary integer programming, and static heuristics. The method demonstrates strong solution quality and illustrating the benefit of learning-driven control within combinatorial optimization for warehouse systems.", "date": "2025-06-03", "categories": []} {"id": "2301.00384", "title": "Correlation Clustering Algorithm for Dynamic Complete Signed Graphs: An Index-based Approach", "abstract": "In this paper, we reduce the complexity of approximating the correlation clustering problem from $O(m\\times\\left( 2+ \\alpha (G) \\right)+n)$ to $O(m+n)$ for any given value of $\\varepsilon$ for a complete signed graph with $n$ vertices and $m$ positive edges where $\\alpha(G)$ is the arboricity of the graph. Our approach gives the same output as the original algorithm and makes it possible to implement the algorithm in a full dynamic setting where edge sign flipping and vertex addition/removal are allowed. Constructing this index costs $O(m)$ memory and $O(m\\times\\alpha(G))$ time. We also studied the structural properties of the non-agreement measure used in the approximation algorithm. The theoretical results are accompanied by a full set of experiments concerning seven real-world graphs. These results shows superiority of our index-based algorithm to the non-index one by a decrease of %34 in time on average.", "date": "2023-01-01", "categories": []} {"id": "2410.05937", "title": "Athanor: Local Search over Abstract Constraint Specifications", "abstract": "Local search is a common method for solving combinatorial optimisation problems. We focus on general-purpose local search solvers that accept as input a constraint model - a declarative description of a problem consisting of a set of decision variables under a set of constraints. Existing approaches typically take as input models written in solver-independent constraint modelling languages like MiniZinc. The Athanor solver we describe herein differs in that it begins from a specification of a problem in the abstract constraint specification language Essence, which allows problems to be described without commitment to low-level modelling decisions through its support for a rich set of abstract types. The advantage of proceeding from Essence is that the structure apparent in a concise, abstract specification of a problem can be exploited to generate high quality neighbourhoods automatically, avoiding the difficult task of identifying that structure in an equivalent constraint model. Based on the twin benefits of neighbourhoods derived from high level types and the scalability derived by searching directly over those types, our empirical results demonstrate strong performance in practice relative to existing solution methods.", "date": "2024-10-08", "categories": []} {"id": "2412.09433", "title": "Solving Multiagent Path Finding on Highly Centralized Networks", "abstract": "The Mutliagent Path Finding (MAPF) problem consists of identifying the trajectories that a set of agents should follow inside a given network in order to reach their desired destinations as soon as possible, but without colliding with each other. We aim to minimize the maximum time any agent takes to reach their goal, ensuring optimal path length. In this work, we complement a recent thread of results that aim to systematically study the algorithmic behavior of this problem, through the parameterized complexity point of view.\n First, we show that MAPF is NP-hard when the given network has a star-like topology (bounded vertex cover number) or is a tree with $11$ leaves. Both of these results fill important gaps in our understanding of the tractability of this problem that were left untreated in the recent work of [Fioravantes et al. Exact Algorithms and Lowerbounds for Multiagent Path Finding: Power of Treelike Topology. AAAI'24]. Nevertheless, our main contribution is an exact algorithm that scales well as the input grows (FPT) when the topology of the given network is highly centralized (bounded distance to clique). This parameter is significant as it mirrors real-world networks. In such environments, a bunch of central hubs (e.g., processing areas) are connected to only few peripheral nodes.", "date": "2024-12-12", "categories": []} {"id": "2503.07580", "title": "BOPO: Neural Combinatorial Optimization via Best-anchored and Objective-guided Preference Optimization", "abstract": "Neural Combinatorial Optimization (NCO) has emerged as a promising approach for NP-hard problems. However, prevailing RL-based methods suffer from low sample efficiency due to sparse rewards and underused solutions. We propose Best-anchored and Objective-guided Preference Optimization (BOPO), a training paradigm that leverages solution preferences via objective values. It introduces: (1) a best-anchored preference pair construction for better explore and exploit solutions, and (2) an objective-guided pairwise loss function that adaptively scales gradients via objective differences, removing reliance on reward models or reference policies. Experiments on Job-shop Scheduling Problem (JSP), Traveling Salesman Problem (TSP), and Flexible Job-shop Scheduling Problem (FJSP) show BOPO outperforms state-of-the-art neural methods, reducing optimality gaps impressively with efficient inference. BOPO is architecture-agnostic, enabling seamless integration with existing NCO models, and establishes preference optimization as a principled framework for combinatorial optimization.", "date": "2025-03-10", "categories": []} {"id": "2505.10328", "title": "A Comparative Study of SMT and MILP for the Nurse Rostering Problem", "abstract": "The effects of personnel scheduling on the quality of care and working conditions for healthcare personnel have been thoroughly documented. However, the ever-present demand and large variation of constraints make healthcare scheduling particularly challenging. This problem has been studied for decades, with limited research aimed at applying Satisfiability Modulo Theories (SMT). SMT has gained momentum within the formal verification community in the last decades, leading to the advancement of SMT solvers that have been shown to outperform standard mathematical programming techniques. In this work, we propose generic constraint formulations that can model a wide range of real-world scheduling constraints. Then, the generic constraints are formulated as SMT and MILP problems and used to compare the respective state-of-the-art solvers, Z3 and Gurobi, on academic and real-world inspired rostering problems. Experimental results show how each solver excels for certain types of problems; the MILP solver generally performs better when the problem is highly constrained or infeasible, while the SMT solver performs better otherwise. On real-world inspired problems containing a more varied set of shifts and personnel, the SMT solver excels. Additionally, it was noted during experimentation that the SMT solver was more sensitive to the way the generic constraints were formulated, requiring careful consideration and experimentation to achieve better performance. We conclude that SMT-based methods present a promising avenue for future research within the domain of personnel scheduling.", "date": "2025-05-15", "categories": []} {"id": "2506.01065", "title": "Trilevel Memetic Algorithm for the Electric Vehicle Routing Problem", "abstract": "The Electric Vehicle Routing Problem (EVRP) extends the capacitated vehicle routing problem by incorporating battery constraints and charging stations, posing significant optimization challenges. This paper introduces a Trilevel Memetic Algorithm (TMA) that hierarchically optimizes customer sequences, route assignments, and charging station insertions. The method combines genetic algorithms with dynamic programming, ensuring efficient and high-quality solutions. Benchmark tests on WCCI2020 instances show competitive performance, matching best-known results for small-scale cases. While computational demands limit scalability, TMA demonstrates strong potential for sustainable logistics planning.", "date": "2025-06-01", "categories": []} {"id": "2412.08556", "title": "Exact Algorithms for Multiagent Path Finding with Communication Constraints on Tree-Like Structures", "abstract": "Consider the scenario where multiple agents have to move in an optimal way through a network, each one towards their ending position while avoiding collisions. By optimal, we mean as fast as possible, which is evaluated by a measure known as the makespan of the proposed solution. This is the setting studied in the Multiagent Path Finding problem. In this work, we additionally provide the agents with a way to communicate with each other. Due to size constraints, it is reasonable to assume that the range of communication of each agent will be limited. What should be the trajectories of the agents to, additionally, maintain a backbone of communication? In this work, we study the Multiagent Path Finding with Communication Constraint problem under the parameterized complexity framework.\n Our main contribution is three exact algorithms that are efficient when considering particular structures for the input network. We provide such algorithms for the case when the communication range and the number of agents (the makespan resp.) are provided in the input and the network has a tree topology, or bounded maximum degree (has a tree-like topology, i.e., bounded treewidth resp.). We complement these results by showing that it is highly unlikely to construct efficient algorithms when considering the number of agents as part of the input, even if the makespan is $3$ and the communication range is $1$.", "date": "2024-12-11", "categories": []} {"id": "2505.23881", "title": "Using Reasoning Models to Generate Search Heuristics that Solve Open Instances of Combinatorial Design Problems", "abstract": "Large Language Models (LLMs) with reasoning are trained to iteratively generate and refine their answers before finalizing them, which can help with applications to mathematics and code generation. We apply code generation with reasoning LLMs to a specific task in the mathematical field of combinatorial design. This field studies diverse types of combinatorial designs, many of which have lists of open instances for which existence has not yet been determined. The Constructive Protocol CPro1 uses LLMs to generate search heuristics that have the potential to construct solutions to small open instances. Starting with a textual definition and a validity verifier for a particular type of design, CPro1 guides LLMs to select and implement strategies, while providing automated hyperparameter tuning and execution feedback. CPro1 with reasoning LLMs successfully solves long-standing open instances for 7 of 16 combinatorial design problems selected from the 2006 Handbook of Combinatorial Designs, including new solved instances for 3 of these (Bhaskar Rao Designs, Symmetric Weighing Matrices, Balanced Ternary Designs) that were unsolved by CPro1 with non-reasoning LLMs. It also solves open instances for several problems from recent (2025) literature, generating new Covering Sequences, Johnson Clique Covers, Deletion Codes, and a Uniform Nested Steiner Quadruple System.", "date": "2025-05-29", "categories": []} {"id": "2505.21887", "title": "SVRPBench: A Realistic Benchmark for Stochastic Vehicle Routing Problem", "abstract": "Robust routing under uncertainty is central to real-world logistics, yet most benchmarks assume static, idealized settings. We present SVRPBench, the first open benchmark to capture high-fidelity stochastic dynamics in vehicle routing at urban scale. Spanning more than 500 instances with up to 1000 customers, it simulates realistic delivery conditions: time-dependent congestion, log-normal delays, probabilistic accidents, and empirically grounded time windows for residential and commercial clients. Our pipeline generates diverse, constraint-rich scenarios, including multi-depot and multi-vehicle setups. Benchmarking reveals that state-of-the-art RL solvers like POMO and AM degrade by over 20% under distributional shift, while classical and metaheuristic methods remain robust. To enable reproducible research, we release the dataset and evaluation suite. SVRPBench challenges the community to design solvers that generalize beyond synthetic assumptions and adapt to real-world uncertainty.", "date": "2025-05-28", "categories": []} {"id": "2505.22753", "title": "Enhancing Lifelong Multi-Agent Path-finding by Using Artificial Potential Fields", "abstract": "We explore the use of Artificial Potential Fields (APFs) to solve Multi-Agent Path Finding (MAPF) and Lifelong MAPF (LMAPF) problems. In MAPF, a team of agents must move to their goal locations without collisions, whereas in LMAPF, new goals are generated upon arrival. We propose methods for incorporating APFs in a range of MAPF algorithms, including Prioritized Planning, MAPF-LNS2, and Priority Inheritance with Backtracking (PIBT). Experimental results show that using APF is not beneficial for MAPF but yields up to a 7-fold increase in overall system throughput for LMAPF.", "date": "2025-05-28", "categories": []} {"id": "2505.23098", "title": "Learning to Search for Vehicle Routing with Multiple Time Windows", "abstract": "In this study, we propose a reinforcement learning-based adaptive variable neighborhood search (RL-AVNS) method designed for effectively solving the Vehicle Routing Problem with Multiple Time Windows (VRPMTW). Unlike traditional adaptive approaches that rely solely on historical operator performance, our method integrates a reinforcement learning framework to dynamically select neighborhood operators based on real-time solution states and learned experience. We introduce a fitness metric that quantifies customers' temporal flexibility to improve the shaking phase, and employ a transformer-based neural policy network to intelligently guide operator selection during the local search. Extensive computational experiments are conducted on realistic scenarios derived from the replenishment of unmanned vending machines, characterized by multiple clustered replenishment windows. Results demonstrate that RL-AVNS significantly outperforms traditional variable neighborhood search (VNS), adaptive VNS (AVNS), and state-of-the-art learning-based heuristics, achieving substantial improvements in solution quality and computational efficiency across various instance scales and time window complexities. Particularly notable is the algorithm's capability to generalize effectively to problem instances not encountered during training, underscoring its practical utility for complex logistics scenarios.", "date": "2025-05-29", "categories": []} {"id": "2505.23584", "title": "Collaborative Last-Mile Delivery: A Multi-Platform Vehicle Routing Problem With En-route Charging", "abstract": "The rapid growth of e-commerce and the increasing demand for timely, cost-effective last-mile delivery have increased interest in collaborative logistics. This research introduces a novel collaborative synchronized multi-platform vehicle routing problem with drones and robots (VRP-DR), where a fleet of $\\mathcal{M}$ trucks, $\\mathcal{N}$ drones and $\\mathcal{K}$ robots, cooperatively delivers parcels. Trucks serve as mobile platforms, enabling the launching, retrieving, and en-route charging of drones and robots, thereby addressing critical limitations such as restricted payload capacities, limited range, and battery constraints. The VRP-DR incorporates five realistic features: (1) multi-visit service per trip, (2) multi-trip operations, (3) flexible docking, allowing returns to the same or different trucks (4) cyclic and acyclic operations, enabling return to the same or different nodes; and (5) en-route charging, enabling drones and robots to recharge while being transported on the truck, maximizing operational efficiency by utilizing idle transit time. The VRP-DR is formulated as a mixed-integer linear program (MILP) to minimize both operational costs and makespan. To overcome the computational challenges of solving large-scale instances, a scalable heuristic algorithm, FINDER (Flexible INtegrated Delivery with Energy Recharge), is developed, to provide efficient, near-optimal solutions. Numerical experiments across various instance sizes evaluate the performance of the MILP and heuristic approaches in terms of solution quality and computation time. The results demonstrate significant time savings of the combined delivery mode over the truck-only mode and substantial cost reductions from enabling multi-visits. The study also provides insights into the effects of en-route charging, docking flexibility, drone count, speed, and payload capacity on system performance.", "date": "2025-05-29", "categories": []} {"id": "2407.10382", "title": "Communication- and Computation-Efficient Distributed Submodular Optimization in Robot Mesh Networks", "abstract": "We provide a communication- and computation-efficient method for distributed submodular optimization in robot mesh networks. Submodularity is a property of diminishing returns that arises in active information gathering such as mapping, surveillance, and target tracking. Our method, Resource-Aware distributed Greedy (RAG), introduces a new distributed optimization paradigm that enables scalable and near-optimal action coordination. To this end, RAG requires each robot to make decisions based only on information received from and about their neighbors. In contrast, the current paradigms allow the relay of information about all robots across the network. As a result, RAG's decision-time scales linearly with the network size, while state-of-the-art near-optimal submodular optimization algorithms scale cubically. We also characterize how the designed mesh-network topology affects RAG's approximation performance. Our analysis implies that sparser networks favor scalability without proportionally compromising approximation performance: while RAG's decision time scales linearly with network size, the gain in approximation performance scales sublinearly. We demonstrate RAG's performance in simulated scenarios of area detection with up to 45 robots, simulating realistic robot-to-robot (r2r) communication speeds such as the 0.25 Mbps speed of the Digi XBee 3 Zigbee 3.0. In the simulations, RAG enables real-time planning, up to three orders of magnitude faster than competitive near-optimal algorithms, while also achieving superior mean coverage performance. To enable the simulations, we extend the high-fidelity and photo-realistic simulator AirSim by integrating a scalable collaborative autonomy pipeline to tens of robots and simulating r2r communication delays. Our code is available at https://github.com/UM-iRaL/Resource-Aware-Coordination-AirSim.", "date": "2024-07-15", "categories": []} {"id": "2505.22384", "title": "Exact Algorithms and Lower Bounds for Forming Coalitions of Constrained Maximum Size", "abstract": "Imagine we want to split a group of agents into teams in the most \\emph{efficient} way, considering that each agent has their own preferences about their teammates. This scenario is modeled by the extensively studied \\textsc{Coalition Formation} problem. Here, we study a version of this problem where each team must additionally be of bounded size.\n We conduct a systematic algorithmic study, providing several intractability results as well as multiple exact algorithms that scale well as the input grows (FPT), which could prove useful in practice.\n Our main contribution is an algorithm that deals efficiently with tree-like structures (bounded \\emph{treewidth}) for ``small'' teams. We complement this result by proving that our algorithm is asymptotically optimal. Particularly, there can be no algorithm that vastly outperforms the one we present, under reasonable theoretical assumptions, even when considering star-like structures (bounded \\emph{vertex cover number}).", "date": "2025-05-28", "categories": []} {"id": "2505.20972", "title": "Deep k-grouping: An Unsupervised Learning Framework for Combinatorial Optimization on Graphs and Hypergraphs", "abstract": "Along with AI computing shining in scientific discovery, its potential in the combinatorial optimization (CO) domain has also emerged in recent years. Yet, existing unsupervised neural network solvers struggle to solve $k$-grouping problems (e.g., coloring, partitioning) on large-scale graphs and hypergraphs, due to limited computational frameworks. In this work, we propose Deep $k$-grouping, an unsupervised learning-based CO framework. Specifically, we contribute: Novel one-hot encoded polynomial unconstrained binary optimization (OH-PUBO), a formulation for modeling k-grouping problems on graphs and hypergraphs (e.g., graph/hypergraph coloring and partitioning); GPU-accelerated algorithms for large-scale k-grouping CO problems. Deep $k$-grouping employs the relaxation of large-scale OH-PUBO objectives as differentiable loss functions and trains to optimize them in an unsupervised manner. To ensure scalability, it leverages GPU-accelerated algorithms to unify the training pipeline; A Gini coefficient-based continuous relaxation annealing strategy to enforce discreteness of solutions while preventing convergence to local optima. Experimental results demonstrate that Deep $k$-grouping outperforms existing neural network solvers and classical heuristics such as SCIP and Tabu.", "date": "2025-05-27", "categories": []} {"id": "2505.20997", "title": "BIPNN: Learning to Solve Binary Integer Programming via Hypergraph Neural Networks", "abstract": "Binary (0-1) integer programming (BIP) is pivotal in scientific domains requiring discrete decision-making. As the advance of AI computing, recent works explore neural network-based solvers for integer linear programming (ILP) problems. Yet, they lack scalability for tackling nonlinear challenges. To handle nonlinearities, state-of-the-art Branch-and-Cut solvers employ linear relaxations, leading to exponential growth in auxiliary variables and severe computation limitations. To overcome these limitations, we propose BIPNN (Binary Integer Programming Neural Network), an unsupervised learning framework to solve nonlinear BIP problems via hypergraph neural networks (HyperGNN). Specifically, BIPNN reformulates BIPs-constrained, discrete, and nonlinear (sin, log, exp) optimization problems-into unconstrained, differentiable, and polynomial loss functions. The reformulation stems from the observation of a precise one-to-one mapping between polynomial BIP objectives and hypergraph structures, enabling the unsupervised training of HyperGNN to optimize BIP problems in an end-to-end manner. On this basis, we propose a GPU-accelerated and continuous-annealing-enhanced training pipeline for BIPNN. The pipeline enables BIPNN to optimize large-scale nonlinear terms in BIPs fully in parallel via straightforward gradient descent, thus significantly reducing the training cost while ensuring the generation of discrete, high-quality solutions. Extensive experiments on synthetic and real-world datasets highlight the superiority of our approach.", "date": "2025-05-27", "categories": []} {"id": "2305.06361", "title": "Efficient Training of Multi-task Neural Solver for Combinatorial Optimization", "abstract": "Efficiently training a multi-task neural solver for various combinatorial optimization problems (COPs) has been less studied so far. Naive application of conventional multi-task learning approaches often falls short in delivering a high-quality, unified neural solver. This deficiency primarily stems from the significant computational demands and a lack of adequate consideration for the complexities inherent in COPs. In this paper, we propose a general and efficient training paradigm to deliver a unified combinatorial multi-task neural solver. To this end, we resort to the theoretical loss decomposition for multiple tasks under an encoder-decoder framework, which enables more efficient training via proper bandit task-sampling algorithms through an intra-task influence matrix. By employing theoretically grounded approximations, our method significantly enhances overall performance, regardless of whether it is within constrained training budgets, across equivalent training epochs, or in terms of generalization capabilities, when compared to conventional training schedules. On the real-world datasets of TSPLib and CVRPLib, our method also achieved the best results compared to single task learning and multi-task learning approaches. Additionally, the influence matrix provides empirical evidence supporting common practices in the field of learning to optimize, further substantiating the effectiveness of our approach. Our code is open-sourced and available at https://github.com/LOGO-CUHKSZ/MTL-COP.", "date": "2023-05-10", "categories": []} {"id": "2410.09693", "title": "Neural Solver Selection for Combinatorial Optimization", "abstract": "Machine learning has increasingly been employed to solve NP-hard combinatorial optimization problems, resulting in the emergence of neural solvers that demonstrate remarkable performance, even with minimal domain-specific knowledge. To date, the community has created numerous open-source neural solvers with distinct motivations and inductive biases. While considerable efforts are devoted to designing powerful single solvers, our findings reveal that existing solvers typically demonstrate complementary performance across different problem instances. This suggests that significant improvements could be achieved through effective coordination of neural solvers at the instance level. In this work, we propose the first general framework to coordinate the neural solvers, which involves feature extraction, selection model, and selection strategy, aiming to allocate each instance to the most suitable solvers. To instantiate, we collect several typical neural solvers with state-of-the-art performance as alternatives, and explore various methods for each component of the framework. We evaluated our framework on two extensively studied combinatorial optimization problems, Traveling Salesman Problem (TSP) and Capacitated Vehicle Routing Problem (CVRP). Experimental results show that the proposed framework can effectively distribute instances and the resulting composite solver can achieve significantly better performance (e.g., reduce the optimality gap by 0.88\\% on TSPLIB and 0.71\\% on CVRPLIB) than the best individual neural solver with little extra time cost.", "date": "2024-10-13", "categories": []} {"id": "2502.19377", "title": "Preference-Based Gradient Estimation for ML-Guided Approximate Combinatorial Optimization", "abstract": "Combinatorial optimization (CO) problems arise across a broad spectrum of domains, including medicine, logistics, and manufacturing. While exact solutions are often computationally infeasible, many practical applications require high-quality solutions within a given time budget. To address this, we propose a learning-based approach that enhances existing non-learned approximation algorithms for CO. Specifically, we parameterize these approximation algorithms and train graph neural networks (GNNs) to predict parameter values that yield near-optimal solutions. Our method is trained end-to-end in a self-supervised fashion, using a novel gradient estimation scheme that treats the approximation algorithm as a black box. This approach combines the strengths of learning and traditional algorithms: the GNN learns from data to guide the algorithm toward better solutions, while the approximation algorithm ensures feasibility. We validate our method on two well-known combinatorial optimization problems: the travelling salesman problem (TSP) and the minimum k-cut problem. Our results demonstrate that the proposed approach is competitive with state-of-the-art learned CO solvers.", "date": "2025-02-26", "categories": []} {"id": "2505.19315", "title": "Demand Selection for VRP with Emission Quota", "abstract": "Combinatorial optimization (CO) problems are traditionally addressed using Operations Research (OR) methods, including metaheuristics. In this study, we introduce a demand selection problem for the Vehicle Routing Problem (VRP) with an emission quota, referred to as QVRP. The objective is to minimize the number of omitted deliveries while respecting the pollution quota. We focus on the demand selection part, called Maximum Feasible Vehicle Assignment (MFVA), while the construction of a routing for the VRP instance is solved using classical OR methods. We propose several methods for selecting the packages to omit, both from machine learning (ML) and OR. Our results show that, in this static problem setting, classical OR-based methods consistently outperform ML-based approaches.", "date": "2025-05-25", "categories": []} {"id": "2505.20242", "title": "RedAHD: Reduction-Based End-to-End Automatic Heuristic Design with Large Language Models", "abstract": "Solving NP-hard combinatorial optimization problems (COPs) (e.g., traveling salesman problems (TSPs) and capacitated vehicle routing problems (CVRPs)) in practice traditionally involves handcrafting heuristics or specifying a search space for finding effective heuristics. The main challenges from these approaches, however, are the sheer amount of domain knowledge and implementation efforts required from human experts. Recently, significant progress has been made to address these challenges, particularly by using large language models (LLMs) to design heuristics within some predetermined generalized algorithmic framework (GAF, e.g., ant colony optimization and guided local search) for building key functions/components (e.g., a priori information on how promising it is to include each edge in a solution for TSP and CVRP). Although existing methods leveraging this idea have shown to yield impressive optimization performance, they are not fully end-to-end and still require considerable manual interventions. In this paper, we propose a novel end-to-end framework, named RedAHD, that enables these LLM-based heuristic design methods to operate without the need of GAFs. More specifically, RedAHD employs LLMs to automate the process of reduction, i.e., transforming the COP at hand into similar COPs that are better-understood, from which LLM-based heuristic design methods can design effective heuristics for directly solving the transformed COPs and, in turn, indirectly solving the original COP. Our experimental results, evaluated on six COPs, show that RedAHD is capable of designing heuristics with competitive or improved results over the state-of-the-art methods with minimal human involvement.", "date": "2025-05-26", "categories": []} {"id": "2208.07582", "title": "Deletion Robust Non-Monotone Submodular Maximization over Matroids", "abstract": "Maximizing a submodular function is a fundamental task in machine learning and in this paper we study the deletion robust version of the problem under the classic matroids constraint. Here the goal is to extract a small size summary of the dataset that contains a high value independent set even after an adversary deleted some elements. We present constant-factor approximation algorithms, whose space complexity depends on the rank $k$ of the matroid and the number $d$ of deleted elements. In the centralized setting we present a $(4.597+O(\\varepsilon))$-approximation algorithm with summary size $O( \\frac{k+d}{\\varepsilon^2}\\log \\frac{k}{\\varepsilon})$ that is improved to a $(3.582+O(\\varepsilon))$-approximation with $O(k + \\frac{d}{\\varepsilon^2}\\log \\frac{k}{\\varepsilon})$ summary size when the objective is monotone. In the streaming setting we provide a $(9.435 + O(\\varepsilon))$-approximation algorithm with summary size and memory $O(k + \\frac{d}{\\varepsilon^2}\\log \\frac{k}{\\varepsilon})$; the approximation factor is then improved to $(5.582+O(\\varepsilon))$ in the monotone case.", "date": "2022-08-16", "categories": []} {"id": "2305.19918", "title": "Fully Dynamic Submodular Maximization over Matroids", "abstract": "Maximizing monotone submodular functions under a matroid constraint is a classic algorithmic problem with multiple applications in data mining and machine learning. We study this classic problem in the fully dynamic setting, where elements can be both inserted and deleted in real-time. Our main result is a randomized algorithm that maintains an efficient data structure with an $\\tilde{O}(k^2)$ amortized update time (in the number of additions and deletions) and yields a $4$-approximate solution, where $k$ is the rank of the matroid.", "date": "2023-05-31", "categories": []} {"id": "2505.17443", "title": "Corporate Needs You to Find the Difference: Revisiting Submodular and Supermodular Ratio Optimization Problems", "abstract": "We study the problem of minimizing or maximizing the average value $ f(S)/|S| $ of a submodular or supermodular set function $ f: 2^V \\to \\mathbb{R} $ over non-empty subsets $ S \\subseteq V $. This generalizes classical problems such as Densest Subgraph (DSG), Densest Supermodular Set (DSS), and Submodular Function Minimization (SFM). Motivated by recent applications, we introduce two broad formulations: Unrestricted Sparsest Submodular Set (USSS) and Unrestricted Densest Supermodular Set (UDSS), which allow for negative and non-monotone functions.\n We show that DSS, SFM, USSS, UDSS, and the Minimum Norm Point (MNP) problem are equivalent under strongly polynomial-time reductions, enabling algorithmic crossover. In particular, viewing these through the lens of the MNP in the base polyhedron, we connect Fujishige's theory with dense decomposition, and show that both Fujishige-Wolfe's algorithm and the heuristic \\textsc{SuperGreedy++} act as universal solvers for all these problems, including sub-modular function minimization.\n Theoretically, we explain why \\textsc{SuperGreedy++} is effective beyond DSS, including for tasks like submodular minimization and minimum $ s $-$ t $ cut. Empirically, we test several solvers, including the Fujishige-Wolfe algorithm on over 400 experiments across seven problem types and large-scale real/synthetic datasets. Surprisingly, general-purpose convex and flow-based methods outperform task-specific baselines, demonstrating that with the right framing, general optimization techniques can be both scalable and state-of-the-art for submodular and supermodular ratio problems.", "date": "2025-05-23", "categories": []} {"id": "2505.18043", "title": "Improved Algorithms for Overlapping and Robust Clustering of Edge-Colored Hypergraphs: An LP-Based Combinatorial Approach", "abstract": "Clustering is a fundamental task in both machine learning and data mining. Among various methods, edge-colored clustering (ECC) has emerged as a useful approach for handling categorical data. Given a hypergraph with (hyper)edges labeled by colors, ECC aims to assign vertex colors to minimize the number of edges where the vertex color differs from the edge's color. However, traditional ECC has inherent limitations, as it enforces a nonoverlapping and exhaustive clustering. To tackle these limitations, three versions of ECC have been studied: Local ECC and Global ECC, which allow overlapping clusters, and Robust ECC, which accounts for vertex outliers. For these problems, both linear programming (LP) rounding algorithms and greedy combinatorial algorithms have been proposed. While these LP-rounding algorithms provide high-quality solutions, they demand substantial computation time; the greedy algorithms, on the other hand, run very fast but often compromise solution quality. In this paper, we present an algorithmic framework that combines the strengths of LP with the computational efficiency of combinatorial algorithms. Both experimental and theoretical analyses show that our algorithms efficiently produce high-quality solutions for all three problems: Local, Global, and Robust ECC. We complement our algorithmic contributions with complexity-theoretic inapproximability results and integrality gap bounds, which suggest that significant theoretical improvements are unlikely. Our results also answer two open questions previously raised in the literature.", "date": "2025-05-23", "categories": []} {"id": "2301.11201", "title": "Relative-Interior Solution for the (Incomplete) Linear Assignment Problem with Applications to the Quadratic Assignment Problem", "abstract": "We study the set of optimal solutions of the dual linear programming formulation of the linear assignment problem (LAP) to propose a method for computing a solution from the relative interior of this set. Assuming that an arbitrary dual-optimal solution and an optimal assignment are available (for which many efficient algorithms already exist), our method computes a relative-interior solution in linear time. Since the LAP occurs as a subproblem in the linear programming (LP) relaxation of the quadratic assignment problem (QAP), we employ our method as a new component in the family of dual-ascent algorithms that provide bounds on the optimal value of the QAP. To make our results applicable to the incomplete QAP, which is of interest in practical use-cases, we also provide a linear-time reduction from the incomplete LAP to the complete LAP along with a mapping that preserves optimality and membership in the relative interior. Our experiments on publicly available benchmarks indicate that our approach with relative-interior solution can frequently provide bounds near the optimum of the LP relaxation and its runtime is much lower when compared to a commercial LP solver.", "date": "2023-01-26", "categories": []} {"id": "2505.15839", "title": "Curriculum Learning in Genetic Programming Guided Local Search for Large-scale Vehicle Routing Problems", "abstract": "Manually designing (meta-)heuristics for the Vehicle Routing Problem (VRP) is a challenging task that requires significant domain expertise. Recently, data-driven approaches have emerged as a promising solution, automatically learning heuristics that perform well on training instances and generalize to unseen test cases. Such an approach learns (meta-)heuristics that can perform well on the training instances, expecting it to generalize well on the unseen test instances. A recent method, named GPGLS, uses Genetic Programming (GP) to learn the utility function in Guided Local Search (GLS) and solved large scale VRP effectively. However, the selection of appropriate training instances during the learning process remains an open question, with most existing studies including GPGLS relying on random instance selection. To address this, we propose a novel method, CL-GPGLS, which integrates Curriculum Learning (CL) into GPGLS. Our approach leverages a predefined curriculum to introduce training instances progressively, starting with simpler tasks and gradually increasing complexity, enabling the model to better adapt and optimize for large-scale VRP (LSVRP). Extensive experiments verify the effectiveness of CL-GPGLS, demonstrating significant performance improvements over three baseline methods.", "date": "2025-05-17", "categories": []} {"id": "2505.16053", "title": "Learning from Algorithm Feedback: One-Shot SAT Solver Guidance with GNNs", "abstract": "Boolean Satisfiability (SAT) solvers are foundational to computer science, yet their performance typically hinges on hand-crafted heuristics. This work introduces Reinforcement Learning from Algorithm Feedback (RLAF) as a paradigm for learning to guide SAT solver branching heuristics with Graph Neural Networks (GNNs). Central to our approach is a novel and generic mechanism for injecting inferred variable weights and polarities into the branching heuristics of existing SAT solvers. In a single forward pass, a GNN assigns these parameters to all variables. Casting this one-shot guidance as a reinforcement learning problem lets us train the GNN with off-the-shelf policy-gradient methods, such as GRPO, directly using the solver's computational cost as the sole reward signal. Extensive evaluations demonstrate that RLAF-trained policies significantly reduce the mean solve times of different base solvers across diverse SAT problem distributions, achieving more than a 2x speedup in some cases, while generalizing effectively to larger and harder problems after training. Notably, these policies consistently outperform expert-supervised approaches based on learning handcrafted weighting heuristics, offering a promising path towards data-driven heuristic design in combinatorial optimization.", "date": "2025-05-21", "categories": []} {"id": "2505.16482", "title": "Minimizing the energy depletion in wireless rechargeable sensor networks using bi-level metaheuristic charging schemes", "abstract": "Recently, Wireless Rechargeable Sensor Networks (WRSNs) that leveraged the advantage of wireless energy transfer technology have opened a promising opportunity in solving the limited energy issue. However, an ineffective charging strategy may reduce the charging performance. Although many practical charging algorithms have been introduced, these studies mainly focus on optimizing the charging path with a fully charging approach. This approach may lead to the death of a series of sensors due to their extended charging latency. This paper introduces a novel partial charging approach that follows a bi-level optimized scheme to minimize energy depletion in WRSNs. We aim at optimizing simultaneously two factors: the charging path and time. To accomplish this, we first formulate a mathematical model of the investigated problem. We then propose two approximate algorithms in which the optimization of the charging path and the charging time are considered as the upper and lower level, respectively. The first algorithm combines a Multi-start Local Search method and a Genetic Algorithm to find a solution. The second algorithm adopts a nested approach that utilizes the advantages of the Multitasking and Covariance Matrix Adaptation Evolutionary Strategies. Experimental validations on various network scenarios demonstrate that our proposed algorithms outperform the existing works.", "date": "2025-05-22", "categories": []} {"id": "2311.18662", "title": "TOP-Former: A Multi-Agent Transformer Approach for the Team Orienteering Problem", "abstract": "Route planning for a fleet of vehicles is an important task in applications such as package delivery, surveillance, or transportation, often integrated within larger Intelligent Transportation Systems (ITS). This problem is commonly formulated as a Vehicle Routing Problem (VRP) known as the Team Orienteering Problem (TOP). Existing solvers for this problem primarily rely on either linear programming, which provides accurate solutions but requires computation times that grow with the size of the problem, or heuristic methods, which typically find suboptimal solutions in a shorter time. In this paper, we introduce TOP-Former, a multi-agent route planning neural network designed to efficiently and accurately solve the Team Orienteering Problem. The proposed algorithm is based on a centralized Transformer neural network capable of learning to encode the scenario (modeled as a graph) and analyze the complete context of all agents to deliver fast, precise, and collaborative solutions. Unlike other neural network-based approaches that adopt a more local perspective, TOP-Former is trained to understand the global situation of the vehicle fleet and generate solutions that maximize long-term expected returns. Extensive experiments demonstrate that the presented system outperforms most state-of-the-art methods in terms of both accuracy and computation speed.", "date": "2023-11-30", "categories": []} {"id": "2505.08451", "title": "Adaptive Bias Generalized Rollout Policy Adaptation on the Flexible Job-Shop Scheduling Problem", "abstract": "The Flexible Job-Shop Scheduling Problem (FJSSP) is an NP-hard combinatorial optimization problem, with several application domains, especially for manufacturing purposes. The objective is to efficiently schedule multiple operations on dissimilar machines. These operations are gathered into jobs, and operations pertaining to the same job need to be scheduled sequentially. Different methods have been previously tested to solve this problem, such as Constraint Solving, Tabu Search, Genetic Algorithms, or Monte Carlo Tree Search (MCTS). We propose a novel algorithm derived from the Generalized Nested Rollout Policy Adaptation, developed to solve the FJSSP. We report encouraging experimental results, as our algorithm performs better than other MCTS-based approaches, even if makespans obtained on large instances are still far from known upper bounds.", "date": "2025-05-13", "categories": []} {"id": "2302.14744", "title": "Tight Mixed-Integer Optimization Formulations for Prescriptive Trees", "abstract": "We focus on modeling the relationship between an input feature vector and the predicted outcome of a trained decision tree using mixed-integer optimization. This can be used in many practical applications where a decision tree or tree ensemble is incorporated into an optimization problem to model the predicted outcomes of a decision. We propose tighter mixed-integer optimization formulations than those previously introduced. Existing formulations can be shown to have linear relaxations that have fractional extreme points, even for the simple case of modeling a single decision tree. A formulation we propose, based on a projected union of polyhedra approach, is ideal for a single decision tree. While the formulation is generally not ideal for tree ensembles or if additional constraints are added, it generally has fewer extreme points, leading to a faster time to solve, particularly if the formulation has relatively few trees. However, previous work has shown that formulations based on a binary representation of the feature vector perform well computationally and hence are attractive for use in practical applications. We present multiple approaches to tighten existing formulations with binary vectors, and show that fractional extreme points are removed when there are multiple splits on the same feature. At an extreme, we prove that this results in ideal formulations for tree ensembles modeling a one-dimensional feature vector. Building on this result, we also show via numerical simulations that these additional constraints result in significantly tighter linear relaxations when the feature vector is low dimensional. We also present instances where the time to solve to optimality is significantly improved using these formulations.", "date": "2023-02-28", "categories": []} {"id": "2502.15359", "title": "ARS: Automatic Routing Solver with Large Language Models", "abstract": "Real-world Vehicle Routing Problems (VRPs) are characterized by a variety of practical constraints, making manual solver design both knowledge-intensive and time-consuming. Although there is increasing interest in automating the design of routing algorithms, existing research has explored only a limited array of VRP variants and fails to adequately address the complex and prevalent constraints encountered in real-world situations. To fill this gap, this paper introduces RoutBench, a benchmark of 1,000 VRP variants derived from 24 attributes, for evaluating the effectiveness of automatic routing solvers in addressing complex constraints. Along with RoutBench, we present the Automatic Routing Solver (ARS), which employs Large Language Model (LLM) agents to enhance a backbone algorithm framework by automatically generating constraint-aware heuristic code, based on problem descriptions and several representative constraints selected from a database. Our experiments show that ARS outperforms state-of-the-art LLM-based methods and commonly used solvers, automatically solving 91.67% of common VRPs and achieving at least a 30% improvement across all benchmarks.", "date": "2025-02-21", "categories": []} {"id": "2505.11526", "title": "Code Retrieval for MILP Instance Generation", "abstract": "Mixed-Integer Linear Programming (MILP) is widely used in fields such as scheduling, logistics, and planning. Enhancing the performance of MILP solvers, particularly learning-based solvers, requires substantial amounts of high-quality data. However, existing methods for MILP instance generation typically necessitate training a separate model for each problem class and are computationally intensive when generating new instances. To address these limitations, we reformulate the MILP Instance Generation task as MILP Code Generation task, enabling efficient, flexible, and interpretable instance generation through code. Since MILP instances generated from code can vary significantly in scale, we introduce MILP-EmbedSim, a new similarity metric that accurately measures the similarity between instances of varying sizes within the same problem class. Leveraging this metric, we propose MILP-Retrieval, a pipeline that retrieves generation code from library to produce MILP instances highly similar to target instance. MILP-Retrieval outperforms baselines in both MILP Code Generation and Instance Generation tasks, provides a novel perspective on MILP instance generation and opens new possibilities for learning-based solvers.", "date": "2025-05-11", "categories": []} {"id": "2505.11636", "title": "Generalization Guarantees for Learning Branch-and-Cut Policies in Integer Programming", "abstract": "Mixed-integer programming (MIP) provides a powerful framework for optimization problems, with Branch-and-Cut (B&C) being the predominant algorithm in state-of-the-art solvers. The efficiency of B&C critically depends on heuristic policies for making sequential decisions, including node selection, cut selection, and branching variable selection. While traditional solvers often employ heuristics with manually tuned parameters, recent approaches increasingly leverage machine learning, especially neural networks, to learn these policies directly from data. A key challenge is to understand the theoretical underpinnings of these learned policies, particularly their generalization performance from finite data. This paper establishes rigorous sample complexity bounds for learning B&C policies where the scoring functions guiding each decision step (node, cut, branch) have a certain piecewise polynomial structure. This structure generalizes the linear models that form the most commonly deployed policies in practice and investigated recently in a foundational series of theoretical works by Balcan et al. Such piecewise polynomial policies also cover the neural network architectures (e.g., using ReLU activations) that have been the focal point of contemporary practical studies. Consequently, our theoretical framework closely reflects the models utilized by practitioners investigating machine learning within B&C, offering a unifying perspective relevant to both established theory and modern empirical research in this area. Furthermore, our theory applies to quite general sequential decision making problems beyond B&C.", "date": "2025-05-16", "categories": []} {"id": "2411.09238", "title": "Beyond the Heatmap: A Rigorous Evaluation of Component Impact in MCTS-Based TSP Solvers", "abstract": "The ``Heatmap + Monte Carlo Tree Search (MCTS)'' paradigm has recently emerged as a prominent framework for solving the Travelling Salesman Problem (TSP). While considerable effort has been devoted to enhancing heatmap sophistication through advanced learning models, this paper rigorously examines whether this emphasis is justified, critically assessing the relative impact of heatmap complexity versus MCTS configuration. Our extensive empirical analysis across diverse TSP scales, distributions, and benchmarks reveals two pivotal insights: 1) The configuration of MCTS strategies significantly influences solution quality, underscoring the importance of meticulous tuning to achieve optimal results and enabling valid comparisons among different heatmap methodologies. 2) A rudimentary, parameter-free heatmap based on the intrinsic $k$-nearest neighbor structure of TSP instances, when coupled with an optimally tuned MCTS, can match or surpass the performance of more sophisticated, learned heatmaps, demonstrating robust generalizability on problem scale and distribution shift. To facilitate rigorous and fair evaluations in future research, we introduce a streamlined pipeline for standardized MCTS hyperparameter tuning. Collectively, these findings challenge the prevalent assumption that heatmap complexity is the primary determinant of performance, advocating instead for a balanced integration and comprehensive evaluation of both learning and search components within this paradigm. Our code is available at: https://github.com/LOGO-CUHKSZ/rethink_mcts_tsp.", "date": "2024-11-14", "categories": []} {"id": "2505.10073", "title": "Multi-Robot Task Allocation for Homogeneous Tasks with Collision Avoidance via Spatial Clustering", "abstract": "In this paper, a novel framework is presented that achieves a combined solution based on Multi-Robot Task Allocation (MRTA) and collision avoidance with respect to homogeneous measurement tasks taking place in industrial environments. The spatial clustering we propose offers to simultaneously solve the task allocation problem and deal with collision risks by cutting the workspace into distinguishable operational zones for each robot. To divide task sites and to schedule robot routes within corresponding clusters, we use K-means clustering and the 2-Opt algorithm. The presented framework shows satisfactory performance, where up to 93\\% time reduction (1.24s against 17.62s) with a solution quality improvement of up to 7\\% compared to the best performing method is demonstrated. Our method also completely eliminates collision points that persist in comparative methods in a most significant sense. Theoretical analysis agrees with the claim that spatial partitioning unifies the apparently disjoint tasks allocation and collision avoidance problems under conditions of many identical tasks to be distributed over sparse geographical areas. Ultimately, the findings in this work are of substantial importance for real world applications where both computational efficiency and operation free from collisions is of paramount importance.", "date": "2025-05-15", "categories": []} {"id": "2505.10296", "title": "Optimizing Electric Bus Charging Scheduling with Uncertainties Using Hierarchical Deep Reinforcement Learning", "abstract": "The growing adoption of Electric Buses (EBs) represents a significant step toward sustainable development. By utilizing Internet of Things (IoT) systems, charging stations can autonomously determine charging schedules based on real-time data. However, optimizing EB charging schedules remains a critical challenge due to uncertainties in travel time, energy consumption, and fluctuating electricity prices. Moreover, to address real-world complexities, charging policies must make decisions efficiently across multiple time scales and remain scalable for large EB fleets. In this paper, we propose a Hierarchical Deep Reinforcement Learning (HDRL) approach that reformulates the original Markov Decision Process (MDP) into two augmented MDPs. To solve these MDPs and enable multi-timescale decision-making, we introduce a novel HDRL algorithm, namely Double Actor-Critic Multi-Agent Proximal Policy Optimization Enhancement (DAC-MAPPO-E). Scalability challenges of the Double Actor-Critic (DAC) algorithm for large-scale EB fleets are addressed through enhancements at both decision levels. At the high level, we redesign the decentralized actor network and integrate an attention mechanism to extract relevant global state information for each EB, decreasing the size of neural networks. At the low level, the Multi-Agent Proximal Policy Optimization (MAPPO) algorithm is incorporated into the DAC framework, enabling decentralized and coordinated charging power decisions, reducing computational complexity and enhancing convergence speed. Extensive experiments with real-world data demonstrate the superior performance and scalability of DAC-MAPPO-E in optimizing EB fleet charging schedules.", "date": "2025-05-15", "categories": []} {"id": "2505.10387", "title": "Multi-Agent Path Finding For Large Agents Is Intractable", "abstract": "The multi-agent path finding (MAPF) problem asks to find a set of paths on a graph such that when synchronously following these paths the agents never encounter a conflict. In the most widespread MAPF formulation, the so-called Classical MAPF, the agents sizes are neglected and two types of conflicts are considered: occupying the same vertex or using the same edge at the same time step. Meanwhile in numerous practical applications, e.g. in robotics, taking into account the agents' sizes is vital to ensure that the MAPF solutions can be safely executed. Introducing large agents yields an additional type of conflict arising when one agent follows an edge and its body overlaps with the body of another agent that is actually not using this same edge (e.g. staying still at some distinct vertex of the graph). Until now it was not clear how harder the problem gets when such conflicts are to be considered while planning. Specifically, it was known that Classical MAPF problem on an undirected graph can be solved in polynomial time, however no complete polynomial-time algorithm was presented to solve MAPF with large agents. In this paper we, for the first time, establish that the latter problem is NP-hard and, thus, if P!=NP no polynomial algorithm for it can, unfortunately, be presented. Our proof is based on the prevalent in the field technique of reducing the seminal 3SAT problem (which is known to be an NP-complete problem) to the problem at hand. In particular, for an arbitrary 3SAT formula we procedurally construct a dedicated graph with specific start and goal vertices and show that the given 3SAT formula is satisfiable iff the corresponding path finding instance has a solution.", "date": "2025-05-15", "categories": []} {"id": "2407.21062", "title": "Hybrid Heuristic Algorithms for Adiabatic Quantum Machine Learning Models", "abstract": "Numerous established machine learning models and various neural network architectures can be restructured as Quadratic Unconstrained Binary Optimization (QUBO) problems. A significant challenge in Adiabatic Quantum Machine Learning (AQML) is the computational demand of the training phase. To mitigate this, approximation techniques inspired by quantum annealing, like Simulated Annealing and Multiple Start Tabu Search (MSTS), have been employed to expedite QUBO-based AQML training. This paper introduces a novel hybrid algorithm that incorporates an \"r-flip\" strategy. This strategy is aimed at solving large-scale QUBO problems more effectively, offering better solution quality and lower computational costs compared to existing MSTS methods. The r-flip approach has practical applications in diverse fields, including cross-docking, supply chain management, machine scheduling, and fraud detection. The paper details extensive computational experiments comparing this r-flip enhanced hybrid heuristic against a standard MSTS approach. These tests utilize both standard benchmark problems and three particularly large QUBO instances. The results indicate that the r-flip enhanced method consistently produces high-quality solutions efficiently, operating within practical time constraints.", "date": "2024-07-26", "categories": []} {"id": "2503.16371", "title": "Reinforcement Learning-based Heuristics to Guide Domain-Independent Dynamic Programming", "abstract": "Domain-Independent Dynamic Programming (DIDP) is a state-space search paradigm based on dynamic programming for combinatorial optimization. In its current implementation, DIDP guides the search using user-defined dual bounds. Reinforcement learning (RL) is increasingly being applied to combinatorial optimization problems and shares several key structures with DP, being represented by the Bellman equation and state-based transition systems. We propose using reinforcement learning to obtain a heuristic function to guide the search in DIDP. We develop two RL-based guidance approaches: value-based guidance using Deep Q-Networks and policy-based guidance using Proximal Policy Optimization. Our experiments indicate that RL-based guidance significantly outperforms standard DIDP and problem-specific greedy heuristics with the same number of node expansions. Further, despite longer node evaluation times, RL guidance achieves better run-time performance than standard DIDP on three of four benchmark domains.", "date": "2025-03-20", "categories": []} {"id": "2505.07830", "title": "An Optimized Evacuation Plan for an Active-Shooter Situation Constrained by Network Capacity", "abstract": "A total of more than 3400 public shootings have occurred in the United States between 2016 and 2022. Among these, 25.1% of them took place in an educational institution, 29.4% at the workplace including office buildings, 19.6% in retail store locations, and 13.4% in restaurants and bars. During these critical scenarios, making the right decisions while evacuating can make the difference between life and death. However, emergency evacuation is intensely stressful, which along with the lack of verifiable real-time information may lead to fatal incorrect decisions. To tackle this problem, we developed a multi-route routing optimization algorithm that determines multiple optimal safe routes for each evacuee while accounting for available capacity along the route, thus reducing the threat of crowding and bottlenecking. Overall, our algorithm reduces the total casualties by 34.16% and 53.3%, compared to our previous routing algorithm without capacity constraints and an expert-advised routing strategy respectively. Further, our approach to reduce crowding resulted in an approximate 50% reduction in occupancy in key bottlenecking nodes compared to both of the other evacuation algorithms.", "date": "2025-04-29", "categories": []} {"id": "2505.08025", "title": "PRISM: Complete Online Decentralized Multi-Agent Pathfinding with Rapid Information Sharing using Motion Constraints", "abstract": "We introduce PRISM (Pathfinding with Rapid Information Sharing using Motion Constraints), a decentralized algorithm designed to address the multi-task multi-agent pathfinding (MT-MAPF) problem. PRISM enables large teams of agents to concurrently plan safe and efficient paths for multiple tasks while avoiding collisions. It employs a rapid communication strategy that uses information packets to exchange motion constraint information, enhancing cooperative pathfinding and situational awareness, even in scenarios without direct communication. We prove that PRISM resolves and avoids all deadlock scenarios when possible, a critical challenge in decentralized pathfinding. Empirically, we evaluate PRISM across five environments and 25 random scenarios, benchmarking it against the centralized Conflict-Based Search (CBS) and the decentralized Token Passing with Task Swaps (TPTS) algorithms. PRISM demonstrates scalability and solution quality, supporting 3.4 times more agents than CBS and handling up to 2.5 times more tasks in narrow passage environments than TPTS. Additionally, PRISM matches CBS in solution quality while achieving faster computation times, even under low-connectivity conditions. Its decentralized design reduces the computational burden on individual agents, making it scalable for large environments. These results confirm PRISM's robustness, scalability, and effectiveness in complex and dynamic pathfinding scenarios.", "date": "2025-05-12", "categories": []} {"id": "2505.08735", "title": "Preference Optimization for Combinatorial Optimization Problems", "abstract": "Reinforcement Learning (RL) has emerged as a powerful tool for neural combinatorial optimization, enabling models to learn heuristics that solve complex problems without requiring expert knowledge. Despite significant progress, existing RL approaches face challenges such as diminishing reward signals and inefficient exploration in vast combinatorial action spaces, leading to inefficiency. In this paper, we propose Preference Optimization, a novel method that transforms quantitative reward signals into qualitative preference signals via statistical comparison modeling, emphasizing the superiority among sampled solutions. Methodologically, by reparameterizing the reward function in terms of policy and utilizing preference models, we formulate an entropy-regularized RL objective that aligns the policy directly with preferences while avoiding intractable computations. Furthermore, we integrate local search techniques into the fine-tuning rather than post-processing to generate high-quality preference pairs, helping the policy escape local optima. Empirical results on various benchmarks, such as the Traveling Salesman Problem (TSP), the Capacitated Vehicle Routing Problem (CVRP) and the Flexible Flow Shop Problem (FFSP), demonstrate that our method significantly outperforms existing RL algorithms, achieving superior convergence efficiency and solution quality.", "date": "2025-05-13", "categories": []} {"id": "2505.04558", "title": "Purity Law for Generalizable Neural TSP Solvers", "abstract": "Achieving generalization in neural approaches across different scales and distributions remains a significant challenge for the Traveling Salesman Problem~(TSP). A key obstacle is that neural networks often fail to learn robust principles for identifying universal patterns and deriving optimal solutions from diverse instances. In this paper, we first uncover Purity Law (PuLa), a fundamental structural principle for optimal TSP solutions, defining that edge prevalence grows exponentially with the sparsity of surrounding vertices. Statistically validated across diverse instances, PuLa reveals a consistent bias toward local sparsity in global optima. Building on this insight, we propose Purity Policy Optimization~(PUPO), a novel training paradigm that explicitly aligns characteristics of neural solutions with PuLa during the solution construction process to enhance generalization. Extensive experiments demonstrate that PUPO can be seamlessly integrated with popular neural solvers, significantly enhancing their generalization performance without incurring additional computational overhead during inference.", "date": "2025-05-07", "categories": []} {"id": "2505.06290", "title": "UniCO: Towards a Unified Model for Combinatorial Optimization Problems", "abstract": "Combinatorial Optimization (CO) encompasses a wide range of problems that arise in many real-world scenarios. While significant progress has been made in developing learning-based methods for specialized CO problems, a unified model with a single architecture and parameter set for diverse CO problems remains elusive. Such a model would offer substantial advantages in terms of efficiency and convenience. In this paper, we introduce UniCO, a unified model for solving various CO problems. Inspired by the success of next-token prediction, we frame each problem-solving process as a Markov Decision Process (MDP), tokenize the corresponding sequential trajectory data, and train the model using a transformer backbone. To reduce token length in the trajectory data, we propose a CO-prefix design that aggregates static problem features. To address the heterogeneity of state and action tokens within the MDP, we employ a two-stage self-supervised learning approach. In this approach, a dynamic prediction model is first trained and then serves as a pre-trained model for subsequent policy generation. Experiments across 10 CO problems showcase the versatility of UniCO, emphasizing its ability to generalize to new, unseen problems with minimal fine-tuning, achieving even few-shot or zero-shot performance. Our framework offers a valuable complement to existing neural CO methods that focus on optimizing performance for individual problems.", "date": "2025-05-07", "categories": []} {"id": "2505.04674", "title": "Dynamic Location Search for Identifying Maximum Weighted Independent\n Sets in Complex Networks", "abstract": " While Artificial intelligence (AI), including Generative AI, are effective at\ngenerating high-quality traffic data and optimization solutions in intelligent\ntransportation systems (ITSs), these techniques often demand significant\ntraining time and computational resources, especially in large-scale and\ncomplex scenarios. To address this, we introduce a novel and efficient\nalgorithm for solving the maximum weighted independent set (MWIS) problem,\nwhich can be used to model many ITSs applications, such as traffic signal\ncontrol and vehicle routing. Given the NP-hard nature of the MWIS problem, our\nproposed algorithm, DynLS, incorporates three key innovations to solve it\neffectively. First, it uses a scores-based adaptive vertex perturbation (SAVP)\ntechnique to accelerate convergence, particularly in sparse graphs. Second, it\nincludes a region location mechanism (RLM) to help escape local optima by\ndynamically adjusting the search space. Finally, it employs a novel variable\nneighborhood descent strategy, ComLS, which combines vertex exchange strategies\nwith a reward mechanism to guide the search toward high-quality solutions. Our\nexperimental results demonstrate DynLS's superior performance, consistently\ndelivering high-quality solutions within 1000 seconds. DynLS outperformed five\nleading algorithms across 360 test instances, achieving the best solution for\n350 instances and surpassing the second-best algorithm, Cyclic-Fast, by 177\ninstances. Moreover, DynLS matched Cyclic-Fast's convergence speed,\nhighlighting its efficiency and practicality. This research represents a\nsignificant advancement in heuristic algorithms for the MWIS problem, offering\na promising approach to aid AI techniques in optimizing intelligent\ntransportation systems.\n", "date": "2025-05-07", "categories": []} {"id": "2505.04757", "title": "Primal-dual algorithm for contextual stochastic combinatorial\n optimization", "abstract": " This paper introduces a novel approach to contextual stochastic optimization,\nintegrating operations research and machine learning to address decision-making\nunder uncertainty. Traditional methods often fail to leverage contextual\ninformation, which underscores the necessity for new algorithms. In this study,\nwe utilize neural networks with combinatorial optimization layers to encode\npolicies. Our goal is to minimize the empirical risk, which is estimated from\npast data on uncertain parameters and contexts. To that end, we present a\nsurrogate learning problem and a generic primal-dual algorithm that is\napplicable to various combinatorial settings in stochastic optimization. Our\napproach extends classic Fenchel-Young loss results and introduces a new\nregularization method using sparse perturbations on the distribution simplex.\nThis allows for tractable updates in the original space and can accommodate\ndiverse objective functions. We demonstrate the linear convergence of our\nalgorithm under certain conditions and provide a bound on the non-optimality of\nthe resulting policy in terms of the empirical risk. Experiments on a\ncontextual stochastic minimum weight spanning tree problem show that our\nalgorithm is efficient and scalable, achieving performance comparable to\nimitation learning of solutions computed using an expensive Lagrangian-based\nheuristic.\n", "date": "2025-05-07", "categories": []} {"id": "2505.05047", "title": "Neural Pathways to Program Success: Hopfield Networks for PERT Analysis", "abstract": " Project and task scheduling under uncertainty remains a fundamental challenge\nin program and project management, where accurate estimation of task durations\nand dependencies is critical for delivering complex, multi project systems. The\nProgram Evaluation and Review Technique provides a probabilistic framework to\nmodel task variability and critical paths. In this paper, the author presents a\nnovel formulation of PERT scheduling as an energy minimization problem within a\nHopfield neural network architecture. By mapping task start times and\nprecedence constraints into a neural computation framework, the networks\ninherent optimization dynamics is exploited to approximate globally consistent\nschedules. The author addresses key theoretical issues related to energy\nfunction differentiability, constraint encoding, and convergence, and extends\nthe Hopfield model for structured precedence graphs. Numerical simulations on\nsynthetic project networks comprising up to 1000 tasks demonstrate the\nviability of this approach, achieving near optimal makespans with minimal\nconstraint violations. The findings suggest that neural optimization models\noffer a promising direction for scalable and adaptive project tasks scheduling\nunder uncertainty in areas such as the agentic AI workflows, microservice based\napplications that the modern AI systems are being built upon.\n", "date": "2025-05-08", "categories": []} {"id": "2505.02963", "title": "Single-Sample and Robust Online Resource Allocation", "abstract": " Online Resource Allocation problem is a central problem in many areas of\nComputer Science, Operations Research, and Economics. In this problem, we\nsequentially receive $n$ stochastic requests for $m$ kinds of shared resources,\nwhere each request can be satisfied in multiple ways, consuming different\namounts of resources and generating different values. The goal is to achieve a\n$(1-\\epsilon)$-approximation to the hindsight optimum, where $\\epsilon>0$ is a\nsmall constant, assuming each resource has a large budget.\n In this paper, we investigate the learnability and robustness of online\nresource allocation. Our primary contribution is a novel Exponential Pricing\nalgorithm with the following properties: 1. It requires only a \\emph{single\nsample} from each of the $n$ request distributions to achieve a\n$(1-\\epsilon)$-approximation for online resource allocation with large budgets.\nSuch an algorithm was previously unknown, even with access to polynomially many\nsamples, as prior work either assumed full distributional knowledge or was\nlimited to i.i.d.\\,or random-order arrivals. 2. It is robust to corruptions in\nthe outliers model and the value augmentation model. Specifically, it maintains\nits $(1 - \\epsilon)$-approximation guarantee under both these robustness\nmodels, resolving the open question posed in Argue, Gupta, Molinaro, and Singla\n(SODA'22). 3. It operates as a simple item-pricing algorithm that ensures\nincentive compatibility.\n The intuition behind our Exponential Pricing algorithm is that the price of a\nresource should adjust exponentially as it is overused or underused. It differs\nfrom conventional approaches that use an online learning algorithm for item\npricing. This departure guarantees that the algorithm will never run out of any\nresource, but loses the usual no-regret properties of online learning\nalgorithms, necessitating a new analytical approach.\n", "date": "2025-05-05", "categories": []} {"id": "2505.00803", "title": "To Repair or Not to Repair? Investigating the Importance of AB-Cycles\n for the State-of-the-Art TSP Heuristic EAX", "abstract": " The Edge Assembly Crossover (EAX) algorithm is the state-of-the-art heuristic\nfor solving the Traveling Salesperson Problem (TSP). It regularly outperforms\nother methods, such as the Lin-Kernighan-Helsgaun heuristic (LKH), across\ndiverse sets of TSP instances. Essentially, EAX employs a two-stage mechanism\nthat focuses on improving the current solutions, first, at the local and,\nsubsequently, at the global level. Although the second phase of the algorithm\nhas been thoroughly studied, configured, and refined in the past, in\nparticular, its first stage has hardly been examined.\n In this paper, we thus focus on the first stage of EAX and introduce a novel\nmethod that quickly verifies whether the AB-cycles, generated during its\ninternal optimization procedure, yield valid tours -- or whether they need to\nbe repaired. Knowledge of the latter is also particularly relevant before\napplying other powerful crossover operators such as the Generalized Partition\nCrossover (GPX). Based on our insights, we propose and evaluate several\nimproved versions of EAX. According to our benchmark study across 10 000\ndifferent TSP instances, the most promising of our proposed EAX variants\ndemonstrates improved computational efficiency and solution quality on\npreviously rather difficult instances compared to the current state-of-the-art\nEAX algorithm.\n", "date": "2025-05-01", "categories": []} {"id": "2505.00561", "title": "Learning to Learn with Quantum Optimization via Quantum Neural Networks", "abstract": " Quantum Approximate Optimization Algorithms (QAOA) promise efficient\nsolutions to classically intractable combinatorial optimization problems by\nharnessing shallow-depth quantum circuits. Yet, their performance and\nscalability often hinge on effective parameter optimization, which remains\nnontrivial due to rugged energy landscapes and hardware noise. In this work, we\nintroduce a quantum meta-learning framework that combines quantum neural\nnetworks, specifically Quantum Long Short-Term Memory (QLSTM) architectures,\nwith QAOA. By training the QLSTM optimizer on smaller graph instances, our\napproach rapidly generalizes to larger, more complex problems, substantially\nreducing the number of iterations required for convergence. Through\ncomprehensive benchmarks on Max-Cut and Sherrington-Kirkpatrick model\ninstances, we demonstrate that QLSTM-based optimizers converge faster and\nachieve higher approximation ratios compared to classical baselines, thereby\noffering a robust pathway toward scalable quantum optimization in the NISQ era.\n", "date": "2025-05-01", "categories": []} {"id": "2504.21135", "title": "QAOA Parameter Transferability for Maximum Independent Set using Graph\n Attention Networks", "abstract": " The quantum approximate optimization algorithm (QAOA) is one of the promising\nvariational approaches of quantum computing to solve combinatorial optimization\nproblems. In QAOA, variational parameters need to be optimized by solving a\nseries of nonlinear, nonconvex optimization programs. In this work, we propose\na QAOA parameter transfer scheme using Graph Attention Networks (GAT) to solve\nMaximum Independent Set (MIS) problems. We prepare optimized parameters for\ngraphs of 12 and 14 vertices and use GATs to transfer their parameters to\nlarger graphs. Additionally, we design a hybrid distributed resource-aware\nalgorithm for MIS (HyDRA-MIS), which decomposes large problems into smaller\nones that can fit onto noisy intermediate-scale quantum (NISQ) computers. We\nintegrate our GAT-based parameter transfer approach to HyDRA-MIS and\ndemonstrate competitive results compared to KaMIS, a state-of-the-art classical\nMIS solver, on graphs with several thousands vertices.\n", "date": "2025-04-29", "categories": []} {"id": "2504.18953", "title": "Application of the Brain Drain Optimization Algorithm to the N-Queens\n Problem", "abstract": " This paper introduces the application of the Brain Drain Optimization\nalgorithm -- a swarm-based metaheuristic inspired by the emigration of\nintellectual elites -- to the N-Queens problem. The N-Queens problem, a classic\ncombinatorial optimization problem, serves as a challenge for applying the\nBRADO. A designed cost function guides the search, and the configurations are\ntuned using a TOPSIS-based multicriteria decision making process. BRADO\nconsistently outperforms alternatives in terms of solution quality, achieving\nfewer threats and better objective function values. To assess BRADO's efficacy,\nit is benchmarked against several established metaheuristic algorithms,\nincluding Particle Swarm Optimization (PSO), Genetic Algorithm (GA),\nImperialist Competitive Algorithm (ICA), Iterated Local Search (ILS), and basic\nLocal Search (LS). The study highlights BRADO's potential as a general-purpose\nsolver for combinatorial problems, opening pathways for future applications in\nother domains of artificial intelligence.\n", "date": "2025-04-26", "categories": []} {"id": "2406.00415", "title": "Neural Combinatorial Optimization Algorithms for Solving Vehicle Routing\n Problems: A Comprehensive Survey with Perspectives", "abstract": " Although several surveys on Neural Combinatorial Optimization (NCO) solvers\nspecifically designed to solve Vehicle Routing Problems (VRPs) have been\nconducted, they did not cover the state-of-the-art (SOTA) NCO solvers emerged\nrecently. More importantly, to establish a comprehensive and up-to-date\ntaxonomy of NCO solvers, we systematically review relevant publications and\npreprints, categorizing them into four distinct types, namely Learning to\nConstruct, Learning to Improve, Learning to Predict-Once, and Learning to\nPredict-Multiplicity solvers. Subsequently, we present the inadequacies of the\nSOTA solvers, including poor generalization, incapability to solve large-scale\nVRPs, inability to address most types of VRP variants simultaneously, and\ndifficulty in comparing these NCO solvers with the conventional Operations\nResearch algorithms. Simultaneously, we discuss on-going efforts, identify open\ninadequacies, as well as propose promising and viable directions to overcome\nthese inadequacies. Notably, existing efforts focus on only one or two of these\ninadequacies, with none attempting to address all of them concurrently. In\naddition, we compare the performance of representative NCO solvers from the\nReinforcement, Supervised, and Unsupervised Learning paradigms across VRPs of\nvarying scales. Finally, following the proposed taxonomy, we provide an\naccompanying web page as a live repository for NCO solvers. Through this survey\nand the live repository, we aim to foster further advancements in the NCO\ncommunity.\n", "date": "2024-06-01", "categories": []} {"id": "2208.11985", "title": "Learning to Prune Instances of Steiner Tree Problem in Graphs", "abstract": " We consider the Steiner tree problem on graphs where we are given a set of\nnodes and the goal is to find a tree sub-graph of minimum weight that contains\nall nodes in the given set, potentially including additional nodes. This is a\nclassical NP-hard combinatorial optimisation problem. In recent years, a\nmachine learning framework called learning-to-prune has been successfully used\nfor solving a diverse range of combinatorial optimisation problems. In this\npaper, we use this learning framework on the Steiner tree problem and show that\neven on this problem, the learning-to-prune framework results in computing\nnear-optimal solutions at a fraction of the time required by commercial ILP\nsolvers. Our results underscore the potential of the learning-to-prune\nframework in solving various combinatorial optimisation problems.\n", "date": "2022-08-25", "categories": []} {"id": "2504.16755", "title": "QAOA-PCA: Enhancing Efficiency in the Quantum Approximate Optimization\n Algorithm via Principal Component Analysis", "abstract": " The Quantum Approximate Optimization Algorithm (QAOA) is a promising\nvariational algorithm for solving combinatorial optimization problems on\nnear-term devices. However, as the number of layers in a QAOA circuit\nincreases, which is correlated with the quality of the solution, the number of\nparameters to optimize grows linearly. This results in more iterations required\nby the classical optimizer, which results in an increasing computational burden\nas more circuit executions are needed. To mitigate this issue, we introduce\nQAOA-PCA, a novel reparameterization technique that employs Principal Component\nAnalysis (PCA) to reduce the dimensionality of the QAOA parameter space. By\nextracting principal components from optimized parameters of smaller problem\ninstances, QAOA-PCA facilitates efficient optimization with fewer parameters on\nlarger instances. Our empirical evaluation on the prominent MaxCut problem\ndemonstrates that QAOA-PCA consistently requires fewer iterations than standard\nQAOA, achieving substantial efficiency gains. While this comes at the cost of a\nslight reduction in approximation ratio compared to QAOA with the same number\nof layers, QAOA-PCA almost always outperforms standard QAOA when matched by\nparameter count. QAOA-PCA strikes a favorable balance between efficiency and\nperformance, reducing optimization overhead without significantly compromising\nsolution quality.\n", "date": "2025-04-23", "categories": []} {"id": "2504.16837", "title": "Approximating Optimal Labelings for Temporal Connectivity", "abstract": " In a temporal graph the edge set dynamically changes over time according to a\nset of time-labels associated with each edge that indicates at which time-steps\nthe edge is available. Two vertices are connected if there is a path connecting\nthem in which the edges are traversed in increasing order of their labels. We\nstudy the problem of scheduling the availability time of the edges of a\ntemporal graph in such a way that all pairs of vertices are connected within a\ngiven maximum allowed time $a$ and the overall number of labels is minimized.\n The problem, known as \\emph{Minimum Aged Labeling} (MAL), has several\napplications in logistics, distribution scheduling, and information spreading\nin social networks, where carefully choosing the time-labels can significantly\nreduce infrastructure costs, fuel consumption, or greenhouse gases.\n The problem MAL has previously been proved to be NP-complete on undirected\ngraphs and \\APX-hard on directed graphs. In this paper, we extend our knowledge\non the complexity and approximability of MAL in several directions. We first\nshow that the problem cannot be approximated within a factor better than\n$O(\\log n)$ when $a\\geq 2$, unless $\\text{P} = \\text{NP}$, and a factor better\nthan $2^{\\log ^{1-\\epsilon} n}$ when $a\\geq 3$, unless $\\text{NP}\\subseteq\n\\text{DTIME}(2^{\\text{polylog}(n)})$, where $n$ is the number of vertices in\nthe graph. Then we give a set of approximation algorithms that, under some\nconditions, almost match these lower bounds. In particular, we show that the\napproximation depends on a relation between $a$ and the diameter of the input\ngraph.\n We further establish a connection with a foundational optimization problem on\nstatic graphs called \\emph{Diameter Constrained Spanning Subgraph} (DCSS) and\nshow that our hardness results also apply to DCSS.\n", "date": "2025-04-23", "categories": []} {"id": "2504.09046", "title": "An Enhanced Iterative Deepening Search Algorithm for the Unrestricted\n Container Rehandling Problem", "abstract": " In container terminal yards, the Container Rehandling Problem (CRP) involves\nrearranging containers between stacks under specific operational rules, and it\nis a pivotal optimization challenge in intelligent container scheduling\nsystems. Existing CRP studies primarily focus on minimizing reallocation costs\nusing two-dimensional bay structures, considering factors such as container\nsize, weight, arrival sequences, and retrieval priorities. This paper\nintroduces an enhanced deepening search algorithm integrated with improved\nlower bounds to boost search efficiency. To further reduce the search space, we\ndesign mutually consistent pruning rules to avoid excessive computational\noverhead. The proposed algorithm is validated on three widely used benchmark\ndatasets for the Unrestricted Container Rehandling Problem (UCRP). Experimental\nresults demonstrate that our approach outperforms state-of-the-art exact\nalgorithms in solving the more general UCRP variant, particularly exhibiting\nsuperior efficiency when handling containers within the same priority group\nunder strict time constraints.\n", "date": "2025-04-12", "categories": []} {"id": "2504.08401", "title": "Graph Reduction with Unsupervised Learning in Column Generation: A\n Routing Application", "abstract": " Column Generation (CG) is a popular method dedicated to enhancing\ncomputational efficiency in large scale Combinatorial Optimization (CO)\nproblems. It reduces the number of decision variables in a problem by solving a\npricing problem. For many CO problems, the pricing problem is an Elementary\nShortest Path Problem with Resource Constraints (ESPPRC). Large ESPPRC\ninstances are difficult to solve to near-optimality. Consequently, we use a\nGraph neural Network (GNN) to reduces the size of the ESPPRC such that it\nbecomes computationally tractable with standard solving techniques. Our GNN is\ntrained by Unsupervised Learning and outputs a distribution for the arcs to be\nretained in the reduced PP. The reduced PP is solved by a local search that\nfinds columns with large reduced costs and speeds up convergence. We apply our\nmethod on a set of Capacitated Vehicle Routing Problems with Time Windows and\nshow significant improvements in convergence compared to simple reduction\ntechniques from the literature. For a fixed computational budget, we improve\nthe objective values by over 9\\% for larger instances. We also analyze the\nperformance of our CG algorithm and test the generalization of our method to\ndifferent classes of instances than the training data.\n", "date": "2025-04-11", "categories": []} {"id": "2412.13888", "title": "Resource Constrained Pathfinding with Enhanced Bidirectional A* Search", "abstract": " The classic Resource Constrained Shortest Path (RCSP) problem aims to find a\ncost optimal path between a pair of nodes in a network such that the resources\nused in the path are within a given limit. Having been studied for over a\ndecade, RCSP has seen recent solutions that utilize heuristic-guided search to\nsolve the constrained problem faster. Building upon the bidirectional A* search\nparadigm, this research introduces a novel constrained search framework that\nuses efficient pruning strategies to allow for accelerated and effective RCSP\nsearch in large-scale networks. Results show that, compared to the state of the\nart, our enhanced framework can significantly reduce the constrained search\ntime, achieving speed-ups of over to two orders of magnitude.\n", "date": "2024-12-18", "categories": []} {"id": "2504.11864", "title": "Moving between high-quality optima using multi-satisfiability\n characteristics in hard-to-solve Max3Sat instances", "abstract": " Gray-box optimization proposes effective and efficient optimizers of general\nuse. To this end, it leverages information about variable dependencies and the\nsubfunction-based problem representation. These approaches were already shown\neffective by enabling \\textit{tunnelling} between local optima even if these\nmoves require the modification of many dependent variables. Tunnelling is\nuseful in solving the maximum satisfiability problem (MaxSat), which can be\nreformulated to Max3Sat. Since many real-world problems can be brought to\nsolving the MaxSat/Max3Sat instances, it is important to solve them effectively\nand efficiently. Therefore, we focus on Max3Sat instances for which tunnelling\nfails to introduce improving moves between locally optimal high-quality\nsolutions and the region of globally optimal solutions. We analyze the features\nof such instances on the ground of phase transitions. Based on these\nobservations, we propose manipulating clause-satisfiability characteristics\nthat allow connecting high-quality solutions distant in the solution space. We\nutilize multi-satisfiability characteristics in the optimizer built from\ntypical gray-box mechanisms. The experimental study shows that the proposed\noptimizer can solve those Max3Sat instances that are out of the grasp of\nstate-of-the-art gray-box optimizers. At the same time, it remains effective\nfor instances that have already been successfully solved by gray-box.\n", "date": "2025-04-16", "categories": []} {"id": "2504.11885", "title": "HyperSAT: Unsupervised Hypergraph Neural Networks for Weighted MaxSAT\n Problems", "abstract": " Graph neural networks (GNNs) have shown promising performance in solving both\nBoolean satisfiability (SAT) and Maximum Satisfiability (MaxSAT) problems due\nto their ability to efficiently model and capture the structural dependencies\nbetween literals and clauses. However, GNN methods for solving Weighted MaxSAT\nproblems remain underdeveloped. The challenges arise from the non-linear\ndependency and sensitive objective function, which are caused by the\nnon-uniform distribution of weights across clauses. In this paper, we present\nHyperSAT, a novel neural approach that employs an unsupervised hypergraph\nneural network model to solve Weighted MaxSAT problems. We propose a hypergraph\nrepresentation for Weighted MaxSAT instances and design a cross-attention\nmechanism along with a shared representation constraint loss function to\ncapture the logical interactions between positive and negative literal nodes in\nthe hypergraph. Extensive experiments on various Weighted MaxSAT datasets\ndemonstrate that HyperSAT achieves better performance than state-of-the-art\ncompetitors.\n", "date": "2025-04-16", "categories": []} {"id": "2411.16667", "title": "OPMOS: Ordered Parallel Algorithm for Multi-Objective Shortest-Paths", "abstract": " The Multi-Objective Shortest-Path (MOS) problem finds a set of Pareto-optimal\nsolutions from a start node to a destination node in a multi-attribute graph.\nThe literature explores multi-objective A*-style algorithmic approaches to\nsolving the NP-hard MOS problem. These approaches use consistent heuristics to\ncompute an exact set of solutions for the goal node. A generalized MOS\nalgorithm maintains a \"frontier\" of partial paths at each node and performs\nordered processing to ensure that Pareto-optimal paths are generated to reach\nthe goal node. The algorithm becomes computationally intractable at a higher\nnumber of objectives due to a rapid increase in the search space for\nnon-dominated paths and the significant increase in Pareto-optimal solutions.\nWhile prior works have focused on algorithmic methods to reduce the complexity,\nwe tackle this challenge by exploiting parallelism to accelerate the MOS\nproblem. The key insight is that MOS algorithms rely on the ordered execution\nof partial paths to maintain high work efficiency. The proposed parallel\nalgorithm (OPMOS) unlocks ordered parallelism and efficiently exploits the\nconcurrent execution of multiple paths in MOS. Experimental evaluation using\nthe NVIDIA GH200 Superchip's 72-core Arm-based CPU shows the performance\nscaling potential of OPMOS on work efficiency and parallelism using a\nreal-world application to ship routing.\n", "date": "2024-11-25", "categories": []} {"id": "2504.08737", "title": "Latency-Aware 2-Opt Monotonic Local Search for Distributed Constraint\n Optimization", "abstract": " Researchers recently extended Distributed Constraint Optimization Problems\n(DCOPs) to Communication-Aware DCOPs so that they are applicable in scenarios\nin which messages can be arbitrarily delayed. Distributed asynchronous local\nsearch and inference algorithms designed for CA-DCOPs are less vulnerable to\nmessage latency than their counterparts for regular DCOPs. However, unlike\nlocal search algorithms for (regular) DCOPs that converge to k-opt solutions\n(with k > 1), that is, they converge to solutions that cannot be improved by a\ngroup of k agents), local search CA-DCOP algorithms are limited to 1-opt\nsolutions only. In this paper, we introduce Latency-Aware Monotonic Distributed\nLocal Search-2 (LAMDLS-2), where agents form pairs and coordinate bilateral\nassignment replacements. LAMDLS-2 is monotonic, converges to a 2-opt solution,\nand is also robust to message latency, making it suitable for CA-DCOPs. Our\nresults indicate that LAMDLS-2 converges faster than MGM-2, a benchmark\nalgorithm, to a similar 2-opt solution, in various message latency scenarios.\n", "date": "2025-02-21", "categories": []} {"id": "2504.09608", "title": "ERL-MPP: Evolutionary Reinforcement Learning with Multi-head Puzzle\n Perception for Solving Large-scale Jigsaw Puzzles of Eroded Gaps", "abstract": " Solving jigsaw puzzles has been extensively studied. While most existing\nmodels focus on solving either small-scale puzzles or puzzles with no gap\nbetween fragments, solving large-scale puzzles with gaps presents distinctive\nchallenges in both image understanding and combinatorial optimization. To\ntackle these challenges, we propose a framework of Evolutionary Reinforcement\nLearning with Multi-head Puzzle Perception (ERL-MPP) to derive a better set of\nswapping actions for solving the puzzles. Specifically, to tackle the\nchallenges of perceiving the puzzle with gaps, a Multi-head Puzzle Perception\nNetwork (MPPN) with a shared encoder is designed, where multiple puzzlet heads\ncomprehensively perceive the local assembly status, and a discriminator head\nprovides a global assessment of the puzzle. To explore the large swapping\naction space efficiently, an Evolutionary Reinforcement Learning (EvoRL) agent\nis designed, where an actor recommends a set of suitable swapping actions from\na large action space based on the perceived puzzle status, a critic updates the\nactor using the estimated rewards and the puzzle status, and an evaluator\ncoupled with evolutionary strategies evolves the actions aligning with the\nhistorical assembly experience. The proposed ERL-MPP is comprehensively\nevaluated on the JPLEG-5 dataset with large gaps and the MIT dataset with\nlarge-scale puzzles. It significantly outperforms all state-of-the-art models\non both datasets.\n", "date": "2025-04-13", "categories": []} {"id": "2412.03332", "title": "On Approximability of $\\ell_2^2$ Min-Sum Clustering", "abstract": " The $\\ell_2^2$ min-sum $k$-clustering problem is to partition an input set\ninto clusters $C_1,\\ldots,C_k$ to minimize $\\sum_{i=1}^k\\sum_{p,q\\in\nC_i}\\|p-q\\|_2^2$. Although $\\ell_2^2$ min-sum $k$-clustering is NP-hard, it is\nnot known whether it is NP-hard to approximate $\\ell_2^2$ min-sum\n$k$-clustering beyond a certain factor.\n In this paper, we give the first hardness-of-approximation result for the\n$\\ell_2^2$ min-sum $k$-clustering problem. We show that it is NP-hard to\napproximate the objective to a factor better than $1.056$ and moreover,\nassuming a balanced variant of the Johnson Coverage Hypothesis, it is NP-hard\nto approximate the objective to a factor better than 1.327.\n We then complement our hardness result by giving a nearly linear time\nparameterized PTAS for $\\ell_2^2$ min-sum $k$-clustering running in time\n$O\\left(n^{1+o(1)}d\\cdot \\exp((k\\cdot\\varepsilon^{-1})^{O(1)})\\right)$, where\n$d$ is the underlying dimension of the input dataset.\n Finally, we consider a learning-augmented setting, where the algorithm has\naccess to an oracle that outputs a label $i\\in[k]$ for input point, thereby\nimplicitly partitioning the input dataset into $k$ clusters that induce an\napproximately optimal solution, up to some amount of adversarial error\n$\\alpha\\in\\left[0,\\frac{1}{2}\\right)$. We give a polynomial-time algorithm that\noutputs a $\\frac{1+\\gamma\\alpha}{(1-\\alpha)^2}$-approximation to $\\ell_2^2$\nmin-sum $k$-clustering, for a fixed constant $\\gamma>0$.\n", "date": "2024-12-04", "categories": []} {"id": "2504.07383", "title": "PROPEL: Supervised and Reinforcement Learning for Large-Scale Supply\n Chain Planning", "abstract": " This paper considers how to fuse Machine Learning (ML) and optimization to\nsolve large-scale Supply Chain Planning (SCP) optimization problems. These\nproblems can be formulated as MIP models which feature both integer\n(non-binary) and continuous variables, as well as flow balance and capacity\nconstraints. This raises fundamental challenges for existing integrations of ML\nand optimization that have focused on binary MIPs and graph problems. To\naddress these, the paper proposes PROPEL, a new framework that combines\noptimization with both supervised and Deep Reinforcement Learning (DRL) to\nreduce the size of search space significantly. PROPEL uses supervised learning,\nnot to predict the values of all integer variables, but to identify the\nvariables that are fixed to zero in the optimal solution, leveraging the\nstructure of SCP applications. PROPEL includes a DRL component that selects\nwhich fixed-at-zero variables must be relaxed to improve solution quality when\nthe supervised learning step does not produce a solution with the desired\noptimality tolerance. PROPEL has been applied to industrial supply chain\nplanning optimizations with millions of variables. The computational results\nshow dramatic improvements in solution times and quality, including a 60%\nreduction in primal integral and an 88% primal gap reduction, and improvement\nfactors of up to 13.57 and 15.92, respectively.\n", "date": "2025-04-10", "categories": []} {"id": "2504.07495", "title": "Bottleneck Identification in Resource-Constrained Project Scheduling via\n Constraint Relaxation", "abstract": " In realistic production scenarios, Advanced Planning and Scheduling (APS)\ntools often require manual intervention by production planners, as the system\nworks with incomplete information, resulting in suboptimal schedules. Often,\nthe preferable solution is not found just because of the too-restrictive\nconstraints specifying the optimization problem, representing bottlenecks in\nthe schedule. To provide computer-assisted support for decision-making, we aim\nto automatically identify bottlenecks in the given schedule while linking them\nto the particular constraints to be relaxed. In this work, we address the\nproblem of reducing the tardiness of a particular project in an obtained\nschedule in the resource-constrained project scheduling problem by relaxing\nconstraints related to identified bottlenecks. We develop two methods for this\npurpose. The first method adapts existing approaches from the job shop\nliterature and utilizes them for so-called untargeted relaxations. The second\nmethod identifies potential improvements in relaxed versions of the problem and\nproposes targeted relaxations. Surprisingly, the untargeted relaxations result\nin improvements comparable to the targeted relaxations.\n", "date": "2025-04-10", "categories": []} {"id": "2305.12352", "title": "Data-driven Mixed Integer Optimization through Probabilistic\n Multi-variable Branching", "abstract": " In this paper, we propose a Pre-trained Mixed Integer Optimization framework\n(PreMIO) that accelerates online mixed integer program (MIP) solving with\noffline datasets and machine learning models. Our method is based on a\ndata-driven multi-variable cardinality branching procedure that splits the MIP\nfeasible region using hyperplanes chosen by the concentration inequalities.\nUnlike most previous ML+MIP approaches that either require complicated\nimplementation or suffer from a lack of theoretical justification, our method\nis simple, flexible, provable, and explainable. Numerical experiments on both\nclassical OR benchmark datasets and real-life instances validate the efficiency\nof our proposed method.\n", "date": "2023-05-21", "categories": []} {"id": "2410.15714", "title": "Offline reinforcement learning for job-shop scheduling problems", "abstract": " Recent advances in deep learning have shown significant potential for solving\ncombinatorial optimization problems in real-time. Unlike traditional methods,\ndeep learning can generate high-quality solutions efficiently, which is crucial\nfor applications like routing and scheduling. However, existing approaches like\ndeep reinforcement learning (RL) and behavioral cloning have notable\nlimitations, with deep RL suffering from slow learning and behavioral cloning\nrelying solely on expert actions, which can lead to generalization issues and\nneglect of the optimization objective. This paper introduces a novel offline RL\nmethod designed for combinatorial optimization problems with complex\nconstraints, where the state is represented as a heterogeneous graph and the\naction space is variable. Our approach encodes actions in edge attributes and\nbalances expected rewards with the imitation of expert solutions. We\ndemonstrate the effectiveness of this method on job-shop scheduling and\nflexible job-shop scheduling benchmarks, achieving superior performance\ncompared to state-of-the-art techniques.\n", "date": "2024-10-21", "categories": []} {"id": "2504.03669", "title": "Self-Learning-Based Optimization for Free-form Pipe Routing in\n Aeroengine with Dynamic Design Environment", "abstract": " Pipe routing is a highly complex, time-consuming, and no-deterministic\npolynomial-time hard (NP-hard) problem in aeroengine design. Despite extensive\nresearch efforts in optimizing constant-curvature pipe routing, the growing\ndemand for free-form pipes poses new challenges. Dynamic design environments\nand fuzzy layout rules further impact the optimization performance and\nefficiency. To tackle these challenges, this study proposes a\nself-learning-based method (SLPR) for optimizing free-form pipe routing in\naeroengines. The SLPR is based on the proximal policy optimization (PPO)\nalgorithm and integrates a unified rule modeling framework for efficient\nobstacle detection and fuzzy rule modeling in continuous space. Additionally, a\npotential energy table is constructed to enable rapid queries of layout\ntendencies and interference. The agent within SLPR iteratively refines pipe\nrouting and accumulates the design knowledge through interaction with the\nenvironment. Once the design environment shifts, the agent can swiftly adapt by\nfine-tuning network parameters. Comparative tests reveal that SLPR ensures\nsmooth pipe routing through cubic non-uniform B-spline (NURBS) curves, avoiding\nredundant pipe segments found in constant-curvature pipe routing. Results in\nboth static and dynamic design environments demonstrate that SLPR outperforms\nthree representative baselines in terms of the pipe length reduction, the\nadherence to layout rules, the path complexity, and the computational\nefficiency. Furthermore, tests in dynamic environments indicate that SLPR\neliminates labor-intensive searches from scratch and even yields superior\nsolutions compared to the retrained model. These results highlight the\npractical value of SLPR for real-world pipe routing, meeting lightweight,\nprecision, and sustainability requirements of the modern aeroengine design.\n", "date": "2025-03-20", "categories": []} {"id": "2504.03688", "title": "CLCR: Contrastive Learning-based Constraint Reordering for Efficient\n MILP Solving", "abstract": " Constraint ordering plays a critical role in the efficiency of Mixed-Integer\nLinear Programming (MILP) solvers, particularly for large-scale problems where\npoorly ordered constraints trigger increased LP iterations and suboptimal\nsearch trajectories. This paper introduces CLCR (Contrastive Learning-based\nConstraint Reordering), a novel framework that systematically optimizes\nconstraint ordering to accelerate MILP solving. CLCR first clusters constraints\nbased on their structural patterns and then employs contrastive learning with a\npointer network to optimize their sequence, preserving problem equivalence\nwhile improving solver efficiency. Experiments on benchmarks show CLCR reduces\nsolving time by 30% and LP iterations by 25% on average, without sacrificing\nsolution accuracy. This work demonstrates the potential of data-driven\nconstraint ordering to enhance optimization models, offering a new paradigm for\nbridging mathematical programming with machine learning.\n", "date": "2025-03-23", "categories": []} {"id": "2504.04874", "title": "Futureproof Static Memory Planning", "abstract": " The NP-complete combinatorial optimization task of assigning offsets to a set\nof buffers with known sizes and lifetimes so as to minimize total memory usage\nis called dynamic storage allocation (DSA). Existing DSA implementations bypass\nthe theoretical state-of-the-art algorithms in favor of either fast but\nwasteful heuristics, or memory-efficient approaches that do not scale beyond\none thousand buffers. The \"AI memory wall\", combined with deep neural networks'\nstatic architecture, has reignited interest in DSA. We present idealloc, a\nlow-fragmentation, high-performance DSA implementation designed for\nmillion-buffer instances. Evaluated on a novel suite of particularly hard\nbenchmarks from several domains, idealloc ranks first against four production\nimplementations in terms of a joint effectiveness/robustness criterion.\n", "date": "2025-04-07", "categories": []} {"id": "2504.03277", "title": "Monte Carlo Graph Coloring", "abstract": " Graph Coloring is probably one of the most studied and famous problem in\ngraph algorithms. Exact methods fail to solve instances with more than few\nhundred vertices, therefore, a large number of heuristics have been proposed.\nNested Monte Carlo Search (NMCS) and Nested Rollout Policy Adaptation (NRPA)\nare Monte Carlo search algorithms for single player games. Surprisingly, few\nwork has been dedicated to evaluating Monte Carlo search algorithms to\ncombinatorial graph problems. In this paper we expose how to efficiently apply\nMonte Carlo search to Graph Coloring and compare this approach to existing\nones.\n", "date": "2025-04-04", "categories": []} {"id": "2402.16442", "title": "On Distributed Larger-Than-Memory Subset Selection With Pairwise\n Submodular Functions", "abstract": " Modern datasets span billions of samples, making training on all available\ndata infeasible. Selecting a high quality subset helps in reducing training\ncosts and enhancing model quality. Submodularity, a discrete analogue of\nconvexity, is commonly used for solving such subset selection problems.\nHowever, existing algorithms for optimizing submodular functions are\nsequential, and the prior distributed methods require at least one central\nmachine to fit the target subset in DRAM. At billion datapoint scale, even the\nsubset may not fit a single machine, and the sequential algorithms are\nprohibitively slow. In this paper, we relax the requirement of having a central\nmachine for the target subset by proposing a novel distributed bounding\nalgorithm with provable approximation guarantees. The algorithm iteratively\nbounds the minimum and maximum utility values to select high quality points and\ndiscard the unimportant ones. When bounding does not find the complete subset,\nwe use a multi-round, partition-based distributed greedy algorithm to identify\nthe remaining subset. We discuss how to implement these algorithms in a\ndistributed data processing framework and empirically analyze different\nconfigurations. We find high quality subsets on CIFAR-100 and ImageNet with\nmarginal or no loss in quality compared to centralized methods, and scale to a\ndataset with 13 billion points.\n", "date": "2024-02-26", "categories": []} {"id": "2502.17725", "title": "Solving the Traveling Salesman Problem via Different Quantum Computing\n Architectures", "abstract": " We study the application of emerging photonic and quantum computing\narchitectures to solving the Traveling Salesman Problem (TSP), a well-known\nNP-hard optimization problem. We investigate several approaches: Simulated\nAnnealing (SA), Quadratic Unconstrained Binary Optimization (QUBO-Ising)\nmethods implemented on quantum annealers and Optical Coherent Ising Machines,\nas well as the Quantum Approximate Optimization Algorithm (QAOA) and the\nQuantum Phase Estimation (QPE) algorithm on gate-based quantum computers. QAOA\nand QPE were tested on the IBM Quantum platform. The QUBO-Ising method was\nexplored using the D-Wave quantum annealer, which operates on superconducting\nJosephson junctions, and the Quantum Computing Inc (QCi) Dirac-1 entropy\nquantum optimization machine. Gate-based quantum computers demonstrated\naccurate results for small TSP instances in simulation. However, real quantum\ndevices are hindered by noise and limited scalability. Circuit complexity grows\nwith problem size, restricting performance to TSP instances with a maximum of 6\nnodes. In contrast, Ising-based architectures show improved scalability for\nlarger problem sizes. SQUID-based Ising machines can handle TSP instances with\nup to 12 nodes, while entropy computing implemented in hybrid optoelectronic\ncomponents extend this capability to 18 nodes. Nevertheless, the solutions tend\nto be suboptimal due to hardware limitations and challenges in achieving ground\nstate convergence as the problem size increases. Despite these limitations,\nIsing machines demonstrate significant time advantages over classical methods,\nmaking them a promising candidate for solving larger-scale TSPs efficiently.\n", "date": "2025-02-24", "categories": []} {"id": "2503.23966", "title": "Machine Learning-assisted High-speed Combinatorial Optimization with\n Ising Machines for Dynamically Changing Problems", "abstract": " Quantum or quantum-inspired Ising machines have recently shown promise in\nsolving combinatorial optimization problems in a short time. Real-world\napplications, such as time division multiple access (TDMA) scheduling for\nwireless multi-hop networks and financial trading, require solving those\nproblems sequentially where the size and characteristics change dynamically.\nHowever, using Ising machines involves challenges to shorten system-wide\nlatency due to the transfer of large Ising model or the cloud access and to\ndetermine the parameters for each problem. Here we show a combinatorial\noptimization method using embedded Ising machines, which enables solving\ndiverse problems at high speed without runtime parameter tuning. We customize\nthe algorithm and circuit architecture of the simulated bifurcation-based Ising\nmachine to compress the Ising model and accelerate computation and then built a\nmachine learning model to estimate appropriate parameters using extensive\ntraining data. In TDMA scheduling for wireless multi-hop networks, our\ndemonstration has shown that the sophisticated system can adapt to changes in\nthe problem and showed that it has a speed advantage over conventional methods.\n", "date": "2025-03-31", "categories": []} {"id": "2504.01173", "title": "Neural Approaches to SAT Solving: Design Choices and Interpretability", "abstract": " In this contribution, we provide a comprehensive evaluation of graph neural\nnetworks applied to Boolean satisfiability problems, accompanied by an\nintuitive explanation of the mechanisms enabling the model to generalize to\ndifferent instances. We introduce several training improvements, particularly a\nnovel closest assignment supervision method that dynamically adapts to the\nmodel's current state, significantly enhancing performance on problems with\nlarger solution spaces. Our experiments demonstrate the suitability of\nvariable-clause graph representations with recurrent neural network updates,\nwhich achieve good accuracy on SAT assignment prediction while reducing\ncomputational demands. We extend the base graph neural network into a diffusion\nmodel that facilitates incremental sampling and can be effectively combined\nwith classical techniques like unit propagation. Through analysis of embedding\nspace patterns and optimization trajectories, we show how these networks\nimplicitly perform a process very similar to continuous relaxations of MaxSAT,\noffering an interpretable view of their reasoning process. This understanding\nguides our design choices and explains the ability of recurrent architectures\nto scale effectively at inference time beyond their training distribution,\nwhich we demonstrate with test-time scaling experiments.\n", "date": "2025-04-01", "categories": []} {"id": "2504.01560", "title": "Optimizing Package Delivery with Quantum Annealers: Addressing\n Time-Windows and Simultaneous Pickup and Delivery", "abstract": " Recent research at the intersection of quantum computing and routing problems\nhas been highly prolific. Much of this work focuses on classical problems such\nas the Traveling Salesman Problem and the Vehicle Routing Problem. The\npractical applicability of these problems depends on the specific objectives\nand constraints considered. However, it is undeniable that translating complex\nreal-world requirements into these classical formulations often proves\nchallenging. In this paper, we resort to our previously published\nquantum-classical technique for addressing real-world-oriented routing\nproblems, known as Quantum for Real Package Delivery (Q4RPD), and elaborate on\nsolving additional realistic problem instances. Accordingly, this paper\nemphasizes the following characteristics: i) simultaneous pickup and\ndeliveries, ii) time-windows, and iii) mobility restrictions by vehicle type.\nTo illustrate the application of Q4RPD, we have conducted an experimentation\ncomprising seven instances, serving as a demonstration of the newly developed\nfeatures.\n", "date": "2025-04-02", "categories": []} {"id": "2410.18832", "title": "MazeNet: An Accurate, Fast, and Scalable Deep Learning Solution for\n Steiner Minimum Trees", "abstract": " The Obstacle Avoiding Rectilinear Steiner Minimum Tree (OARSMT) problem,\nwhich seeks the shortest interconnection of a given number of terminals in a\nrectilinear plane while avoiding obstacles, is a critical task in integrated\ncircuit design, network optimization, and robot path planning. Since OARSMT is\nNP-hard, exact algorithms scale poorly with the number of terminals, leading\npractical solvers to sacrifice accuracy for large problems. We propose MazeNet,\na deep learning-based method that learns to solve the OARSMT from data. MazeNet\nreframes OARSMT as a maze-solving task that can be addressed with a recurrent\nconvolutional neural network (RCNN). A key hallmark of MazeNet is its\nscalability: we only need to train the RCNN blocks on mazes with a small number\nof terminals; larger mazes can be solved by replicating the same pre-trained\nblocks to create a larger network. Across a wide range of experiments, MazeNet\nachieves perfect OARSMT-solving accuracy, significantly reduces runtime\ncompared to classical exact algorithms, and can handle more terminals than\nstate-of-the-art approximate algorithms.\n", "date": "2024-10-24", "categories": []} {"id": "2503.24325", "title": "Pro-Routing: Proactive Routing of Autonomous Multi-Capacity Robots for\n Pickup-and-Delivery Tasks", "abstract": " We consider a multi-robot setting, where we have a fleet of multi-capacity\nautonomous robots that must service spatially distributed pickup-and-delivery\nrequests with fixed maximum wait times. Requests can be either scheduled ahead\nof time or they can enter the system in real-time. In this setting, stability\nfor a routing policy is defined as the cost of the policy being uniformly\nbounded over time. Most previous work either solve the problem offline to\ntheoretically maintain stability or they consider dynamically arriving requests\nat the expense of the theoretical guarantees on stability. In this paper, we\naim to bridge this gap by proposing a novel proactive rollout-based routing\nframework that adapts to real-time demand while still provably maintaining the\nstability of the learned routing policy. We derive provable stability\nguarantees for our method by proposing a fleet sizing algorithm that obtains a\nsufficiently large fleet that ensures stability by construction. To validate\nour theoretical results, we consider a case study on real ride requests for\nHarvard's evening Van System. We also evaluate the performance of our framework\nusing the currently deployed smaller fleet size. In this smaller setup, we\ncompare against the currently deployed routing algorithm, greedy heuristics,\nand Monte-Carlo-Tree-Search-based algorithms. Our empirical results show that\nour framework maintains stability when we use the sufficiently large fleet size\nfound in our theoretical results. For the smaller currently deployed fleet\nsize, our method services 6% more requests than the closest baseline while\nreducing median passenger wait times by 33%.\n", "date": "2025-03-31", "categories": []} {"id": "2503.21814", "title": "Unsupervised Ordering for Maximum Clique", "abstract": " We propose an unsupervised approach for learning vertex orderings for the\nmaximum clique problem by framing it within a permutation-based framework. We\ntransform the combinatorial constraints into geometric relationships such that\nthe ordering of vertices aligns with the clique structures. By integrating this\nclique-oriented ordering into branch-and-bound search, we improve search\nefficiency and reduce the number of computational steps. Our results\ndemonstrate how unsupervised learning of vertex ordering can enhance search\nefficiency across diverse graph instances. We further study the generalization\nacross different sizes.\n", "date": "2025-03-25", "categories": []} {"id": "2503.01877", "title": "Starjob: Dataset for LLM-Driven Job Shop Scheduling", "abstract": " Large Language Models (LLMs) have shown remarkable capabilities across\nvarious domains, but their potential for solving combinatorial optimization\nproblems remains largely unexplored. In this paper, we investigate the\napplicability of LLMs to the Job Shop Scheduling Problem (JSSP), a classic\nchallenge in combinatorial optimization that requires efficient job allocation\nto machines to minimize makespan. To this end, we introduce Starjob, the first\nsupervised dataset for JSSP, comprising 130k instances specifically designed\nfor training LLMs. Leveraging this dataset, we fine-tune the LLaMA 8B 4-bit\nquantized model with the LoRA method to develop an end-to-end scheduling\napproach. Our evaluation on standard benchmarks demonstrates that the proposed\nLLM-based method not only surpasses traditional Priority Dispatching Rules\n(PDRs) but also achieves notable improvements over state-of-the-art neural\napproaches like L2D, with an average improvement of 15.36% on DMU and 7.85% on\nTaillard benchmarks. These results highlight the untapped potential of LLMs in\ntackling combinatorial optimization problems, paving the way for future\nadvancements in this area.\n", "date": "2025-02-26", "categories": []} {"id": "2503.18668", "title": "Geometric Preference Elicitation for Minimax Regret Optimization in\n Uncertainty Matroids", "abstract": " This paper presents an efficient preference elicitation framework for\nuncertain matroid optimization, where precise weight information is\nunavailable, but insights into possible weight values are accessible. The core\ninnovation of our approach lies in its ability to systematically elicit user\npreferences, aligning the optimization process more closely with\ndecision-makers' objectives. By incrementally querying preferences between\npairs of elements, we iteratively refine the parametric uncertainty regions,\nleveraging the structural properties of matroids. Our method aims to achieve\nthe exact optimum by reducing regret with a few elicitation rounds.\nAdditionally, our approach avoids the computation of Minimax Regret and the use\nof Linear programming solvers at every iteration, unlike previous methods.\nExperimental results on four standard matroids demonstrate that our method\nreaches optimality more quickly and with fewer preference queries than existing\ntechniques.\n", "date": "2025-03-24", "categories": []} {"id": "2409.09107", "title": "Proactive and Reactive Constraint Programming for Stochastic Project\n Scheduling with Maximal Time-Lags", "abstract": " This study investigates scheduling strategies for the stochastic\nresource-constrained project scheduling problem with maximal time lags\n(SRCPSP/max)). Recent advances in Constraint Programming (CP) and Temporal\nNetworks have reinvoked interest in evaluating the advantages and drawbacks of\nvarious proactive and reactive scheduling methods. First, we present a new,\nCP-based fully proactive method. Second, we show how a reactive approach can be\nconstructed using an online rescheduling procedure. A third contribution is\nbased on partial order schedules and uses Simple Temporal Networks with\nUncertainty (STNUs). Our statistical analysis shows that the STNU-based\nalgorithm performs best in terms of solution quality, while also showing good\nrelative offline and online computation time.\n", "date": "2024-09-13", "categories": []} {"id": "2503.17573", "title": "Optimizing 2D+1 Packing in Constrained Environments Using Deep\n Reinforcement Learning", "abstract": " This paper proposes a novel approach based on deep reinforcement learning\n(DRL) for the 2D+1 packing problem with spatial constraints. This problem is an\nextension of the traditional 2D packing problem, incorporating an additional\nconstraint on the height dimension. Therefore, a simulator using the OpenAI Gym\nframework has been developed to efficiently simulate the packing of rectangular\npieces onto two boards with height constraints. Furthermore, the simulator\nsupports multidiscrete actions, enabling the selection of a position on either\nboard and the type of piece to place. Finally, two DRL-based methods (Proximal\nPolicy Optimization -- PPO and the Advantage Actor-Critic -- A2C) have been\nemployed to learn a packing strategy and demonstrate its performance compared\nto a well-known heuristic baseline (MaxRect-BL). In the experiments carried\nout, the PPO-based approach proved to be a good solution for solving complex\npackaging problems and highlighted its potential to optimize resource\nutilization in various industrial applications, such as the manufacturing of\naerospace composites.\n", "date": "2025-03-21", "categories": []} {"id": "2501.18432", "title": "Solving Drone Routing Problems with Quantum Computing: A Hybrid Approach\n Combining Quantum Annealing and Gate-Based Paradigms", "abstract": " This paper presents a novel hybrid approach to solving real-world drone\nrouting problems by leveraging the capabilities of quantum computing. The\nproposed method, coined Quantum for Drone Routing (Q4DR), integrates the two\nmost prominent paradigms in the field: quantum gate-based computing, through\nthe Eclipse Qrisp programming language; and quantum annealers, by means of\nD-Wave System's devices. The algorithm is divided into two different phases: an\ninitial clustering phase executed using a Quantum Approximate Optimization\nAlgorithm (QAOA), and a routing phase employing quantum annealers. The efficacy\nof Q4DR is demonstrated through three use cases of increasing complexity, each\nincorporating real-world constraints such as asymmetric costs, forbidden paths,\nand itinerant charging points. This research contributes to the growing body of\nwork in quantum optimization, showcasing the practical applications of quantum\ncomputing in logistics and route planning.\n", "date": "2025-01-30", "categories": []} {"id": "2503.15847", "title": "Beyond Local Selection: Global Cut Selection for Enhanced Mixed-Integer\n Programming", "abstract": " In mixed-integer programming (MIP) solvers, cutting planes are essential for\nBranch-and-Cut (B&C) algorithms as they reduce the search space and accelerate\nthe solving process. Traditional methods rely on hard-coded heuristics for cut\nplane selection but fail to leverage problem-specific structural features.\nRecent machine learning approaches use neural networks for cut selection but\nfocus narrowly on the efficiency of single-node within the B&C algorithm,\nwithout considering the broader contextual information. To address this, we\npropose Global Cut Selection (GCS), which uses a bipartite graph to represent\nthe search tree and combines graph neural networks with reinforcement learning\nto develop cut selection strategies. Unlike prior methods, GCS applies cutting\nplanes across all nodes, incorporating richer contextual information.\nExperiments show GCS significantly improves solving efficiency for synthetic\nand large-scale real-world MIPs compared to traditional and learning-based\nmethods.\n", "date": "2025-03-20", "categories": []} {"id": "2406.06652", "title": "Improving Generalization of Neural Vehicle Routing Problem Solvers\n Through the Lens of Model Architecture", "abstract": " Neural models produce promising results when solving Vehicle Routing Problems\n(VRPs), but often fall short in generalization. Recent attempts to enhance\nmodel generalization often incur unnecessarily large training cost or cannot be\ndirectly applied to other models solving different VRP variants. To address\nthese issues, we take a novel perspective on model architecture in this study.\nSpecifically, we propose a plug-and-play Entropy-based Scaling Factor (ESF) and\na Distribution-Specific (DS) decoder to enhance the size and distribution\ngeneralization, respectively. ESF adjusts the attention weight pattern of the\nmodel towards familiar ones discovered during training when solving VRPs of\nvarying sizes. The DS decoder explicitly models VRPs of multiple training\ndistribution patterns through multiple auxiliary light decoders, expanding the\nmodel representation space to encompass a broader range of distributional\nscenarios. We conduct extensive experiments on both synthetic and widely\nrecognized real-world benchmarking datasets and compare the performance with\nseven baseline models. The results demonstrate the effectiveness of using ESF\nand DS decoder to obtain a more generalizable model and showcase their\napplicability to solve different VRP variants, i.e., travelling salesman\nproblem and capacitated VRP. Notably, our proposed generic components require\nminimal computational resources, and can be effortlessly integrated into\nconventional generalization strategies to further elevate model generalization.\n", "date": "2024-06-10", "categories": []} {"id": "2308.12794", "title": "Job Shop Scheduling Benchmark: Environments and Instances for Learning\n and Non-learning Methods", "abstract": " Job shop scheduling problems address the routing and sequencing of tasks in a\njob shop setting. Despite significant interest from operations research and\nmachine learning communities over the years, a comprehensive platform for\ntesting and comparing solution methods has been notably lacking. To fill this\ngap, we introduce a unified implementation of job shop scheduling problems and\ntheir solution methods, addressing the long-standing need for a standardized\nbenchmarking platform in this domain. Our platform supports classic Job Shop\n(JSP), Flow Shop (FSP), Flexible Job Shop (FJSP), and Assembly Job Shop (AJSP),\nas well as variants featuring Sequence-Dependent Setup Times (SDST), variants\nwith online arrivals of jobs, and combinations of these problems (e.g.,\nFJSP-SDST and FAJSP). The platfrom provides a wide range of scheduling solution\nmethods, from heuristics, metaheuristics, and exact optimization to deep\nreinforcement learning. The implementation is available as an open-source\nGitHub repository, serving as a collaborative hub for researchers,\npractitioners, and those new to the field. Beyond enabling direct comparisons\nwith existing methods on widely studied benchmark problems, this resource\nserves as a robust starting point for addressing constrained and complex\nproblem variants. By establishing a comprehensive and unified foundation, this\nplatform is designed to consolidate existing knowledge and to inspire the\ndevelopment of next-generation algorithms in job shop scheduling research.\n", "date": "2023-08-24", "categories": []} {"id": "2503.11848", "title": "Optimization-Augmented Machine Learning for Vehicle Operations in\n Emergency Medical Services", "abstract": " Minimizing response times to meet legal requirements and serve patients in a\ntimely manner is crucial for Emergency Medical Service (EMS) systems. Achieving\nthis goal necessitates optimizing operational decision-making to efficiently\nmanage ambulances. Against this background, we study a centrally controlled EMS\nsystem for which we learn an online ambulance dispatching and redeployment\npolicy that aims at minimizing the mean response time of ambulances within the\nsystem by dispatching an ambulance upon receiving an emergency call and\nredeploying it to a waiting location upon the completion of its service. We\npropose a novel combinatorial optimization-augmented machine learning pipeline\nthat allows to learn efficient policies for ambulance dispatching and\nredeployment. In this context, we further show how to solve the underlying\nfull-information problem to generate training data and propose an augmentation\nscheme that improves our pipeline's generalization performance by mitigating a\npossible distribution mismatch with respect to the considered state space.\nCompared to existing methods that rely on augmentation during training, our\napproach offers substantial runtime savings of up to 87.9% while yielding\ncompetitive performance. To evaluate the performance of our pipeline against\ncurrent industry practices, we conduct a numerical case study on the example of\nSan Francisco's 911 call data. Results show that the learned policies\noutperform the online benchmarks across various resource and demand scenarios,\nyielding a reduction in mean response time of up to 30%.\n", "date": "2025-03-14", "categories": []} {"id": "2503.10421", "title": "Towards Constraint-Based Adaptive Hypergraph Learning for Solving\n Vehicle Routing: An End-to-End Solution", "abstract": " The application of learning based methods to vehicle routing problems has\nemerged as a pivotal area of research in combinatorial optimization. These\nproblems are characterized by vast solution spaces and intricate constraints,\nmaking traditional approaches such as exact mathematical models or heuristic\nmethods prone to high computational overhead or reliant on the design of\ncomplex heuristic operators to achieve optimal or near optimal solutions.\nMeanwhile, although some recent learning-based methods can produce good\nperformance for VRP with straightforward constraint scenarios, they often fail\nto effectively handle hard constraints that are common in practice. This study\nintroduces a novel end-to-end framework that combines constraint-oriented\nhypergraphs with reinforcement learning to address vehicle routing problems. A\ncentral innovation of this work is the development of a constraint-oriented\ndynamic hyperedge reconstruction strategy within an encoder, which\nsignificantly enhances hypergraph representation learning. Additionally, the\ndecoder leverages a double-pointer attention mechanism to iteratively generate\nsolutions. The proposed model is trained by incorporating asynchronous\nparameter updates informed by hypergraph constraints and optimizing a dual loss\nfunction comprising constraint loss and policy gradient loss. The experiment\nresults on benchmark datasets demonstrate that the proposed approach not only\neliminates the need for sophisticated heuristic operators but also achieves\nsubstantial improvements in solution quality.\n", "date": "2025-03-13", "categories": []} {"id": "2503.08705", "title": "A Block-Based Heuristic Algorithm for the Three-Dimensional Nuclear\n Waste Packing Problem", "abstract": " In this study, we present a block-based heuristic search algorithm to address\nthe nuclear waste container packing problem in the context of real-world\nnuclear power plants. Additionally, we provide a dataset comprising 1600\nproblem instances for future researchers to use. Experimental results on this\ndataset demonstrate that the proposed algorithm effectively enhances the\ndisposal pool's space utilization while minimizing the radiation dose within\nthe pool. The code and data employed in this study are publicly available to\nfacilitate reproducibility and further investigation.\n", "date": "2025-03-09", "categories": []} {"id": "2503.08711", "title": "A Beam Search Based Parallel Algorithm for the Two-Dimensional Strip\n Packing Problem", "abstract": " This paper introduces BSPA, a parallel algorithm that leverages beam search\nto address the two-dimensional strip packing problem. The study begins with a\ncomprehensive review of existing approaches and methodologies, followed by a\ndetailed presentation of the BSPA algorithm. Experimental results demonstrate\nthe effectiveness of the proposed method. To facilitate further research, both\nthe code and datasets are publicly available.\n", "date": "2025-03-10", "categories": []} {"id": "2503.08752", "title": "Large Neighborhood Search and Bitmask Dynamic Programming for Wireless\n Mobile Charging Electric Vehicle Routing Problems in Medical Transportation", "abstract": " The transition to electric vehicles (EVs) is critical to achieving\nsustainable transportation, but challenges such as limited driving range and\ninsufficient charging infrastructure have hindered the widespread adoption of\nEVs, especially in time-sensitive logistics such as medical transportation.\nThis paper presents a new model to break through this barrier by combining\nwireless mobile charging technology with optimization. We propose the Wireless\nMobile Charging Electric Vehicle Routing Problem (WMC-EVRP), which enables\nMedical Transportation Electric Vehicles (MTEVs) to be charged while traveling\nvia Mobile Charging Carts (MCTs). This eliminates the time wastage of stopping\nfor charging and ensures uninterrupted operation of MTEVs for such\ntime-sensitive transportation problems. However, in this problem, the decisions\nof these two types of heterogeneous vehicles are coupled with each other, which\ngreatly increases the difficulty of vehicle routing optimizations. To address\nthis complex problem, we develop a mathematical model and a tailored\nmeta-heuristic algorithm that combines Bit Mask Dynamic Programming (BDP) and\nLarge Neighborhood Search (LNS). The BDP approach efficiently optimizes\ncharging strategies, while the LNS framework utilizes custom operators to\noptimize the MTEV routes under capacity and synchronization constraints. Our\napproach outperforms traditional solvers in providing solutions for medium and\nlarge instances. Using actual hospital locations in Singapore as data, we\nvalidated the practical applicability of the model through extensive\nexperiments and provided important insights into minimizing costs and ensuring\nthe timely delivery of healthcare services.\n", "date": "2025-03-11", "categories": []} {"id": "2503.05963", "title": "Bayesian Graph Traversal", "abstract": " This research considers Bayesian decision-analytic approaches toward the\ntraversal of an uncertain graph. Namely, a traveler progresses over a graph in\nwhich rewards are gained upon a node's first visit and costs are incurred for\nevery edge traversal. The traveler knows the graph's adjacency matrix and his\nstarting position but does not know the rewards and costs. The traveler is a\nBayesian who encodes his beliefs about these values using a Gaussian process\nprior and who seeks to maximize his expected utility over these beliefs.\nAdopting a decision-analytic perspective, we develop sequential decision-making\nsolution strategies for this coupled information-collection and network-routing\nproblem. We show that the problem is NP-Hard and derive properties of the\noptimal walk. These properties provide heuristics for the traveler's problem\nthat balance exploration and exploitation. We provide a practical case study\nfocused on the use of unmanned aerial systems for public safety and empirically\nstudy policy performance in myriad Erdos-Renyi settings.\n", "date": "2025-03-07", "categories": []} {"id": "2503.05071", "title": "Object Packing and Scheduling for Sequential 3D Printing: a Linear\n Arithmetic Model and a CEGAR-inspired Optimal Solver", "abstract": " We address the problem of object arrangement and scheduling for sequential 3D\nprinting. Unlike the standard 3D printing, where all objects are printed slice\nby slice at once, in sequential 3D printing, objects are completed one after\nother. In the sequential case, it is necessary to ensure that the moving parts\nof the printer do not collide with previously printed objects. We look at the\nsequential printing problem from the perspective of combinatorial optimization.\nWe propose to express the problem as a linear arithmetic formula, which is then\nsolved using a solver for satisfiability modulo theories (SMT). However, we do\nnot solve the formula expressing the problem of object arrangement and\nscheduling directly, but we have proposed a technique inspired by\ncounterexample guided abstraction refinement (CEGAR), which turned out to be a\nkey innovation to efficiency.\n", "date": "2025-03-07", "categories": []} {"id": "2503.05175", "title": "Self-Supervised Penalty-Based Learning for Robust Constrained\n Optimization", "abstract": " We propose a new methodology for parameterized constrained robust\noptimization, an important class of optimization problems under uncertainty,\nbased on learning with a self-supervised penalty-based loss function. Whereas\nsupervised learning requires pre-solved instances for training, our approach\nleverages a custom loss function derived from the exact penalty method in\noptimization to learn an approximation, typically defined by a neural network\nmodel, of the parameterized optimal solution mapping. Additionally, we adapt\nour approach to robust constrained combinatorial optimization problems by\nincorporating a surrogate linear cost over mixed integer domains, and a smooth\napproximations thereof, into the final layer of the network architecture. We\nperform computational experiments to test our approach on three different\napplications: multidimensional knapsack with continuous variables,\ncombinatorial multidimensional knapsack with discrete variables, and an\ninventory management problem. Our results demonstrate that our self-supervised\napproach is able to effectively learn neural network approximations whose\ninference time is significantly smaller than the computation time of\ntraditional solvers for this class of robust optimization problems.\nFurthermore, our results demonstrate that by varying the penalty parameter we\nare able to effectively balance the trade-off between sub-optimality and robust\nfeasibility of the obtained solutions.\n", "date": "2025-03-07", "categories": []} {"id": "2503.04047", "title": "Reheated Gradient-based Discrete Sampling for Combinatorial Optimization", "abstract": " Recently, gradient-based discrete sampling has emerged as a highly efficient,\ngeneral-purpose solver for various combinatorial optimization (CO) problems,\nachieving performance comparable to or surpassing the popular data-driven\napproaches. However, we identify a critical issue in these methods, which we\nterm ''wandering in contours''. This behavior refers to sampling new different\nsolutions that share very similar objective values for a long time, leading to\ncomputational inefficiency and suboptimal exploration of potential solutions.\nIn this paper, we introduce a novel reheating mechanism inspired by the concept\nof critical temperature and specific heat in physics, aimed at overcoming this\nlimitation. Empirically, our method demonstrates superiority over existing\nsampling-based and data-driven algorithms across a diverse array of CO\nproblems.\n", "date": "2025-03-06", "categories": []} {"id": "2503.04085", "title": "SED2AM: Solving Multi-Trip Time-Dependent Vehicle Routing Problem using\n Deep Reinforcement Learning", "abstract": " Deep reinforcement learning (DRL)-based frameworks, featuring\nTransformer-style policy networks, have demonstrated their efficacy across\nvarious vehicle routing problem (VRP) variants. However, the application of\nthese methods to the multi-trip time-dependent vehicle routing problem\n(MTTDVRP) with maximum working hours constraints -- a pivotal element of urban\nlogistics -- remains largely unexplored. This paper introduces a DRL-based\nmethod called the Simultaneous Encoder and Dual Decoder Attention Model\n(SED2AM), tailored for the MTTDVRP with maximum working hours constraints. The\nproposed method introduces a temporal locality inductive bias to the encoding\nmodule of the policy networks, enabling it to effectively account for the\ntime-dependency in travel distance or time. The decoding module of SED2AM\nincludes a vehicle selection decoder that selects a vehicle from the fleet,\neffectively associating trips with vehicles for functional multi-trip routing.\nAdditionally, this decoding module is equipped with a trip construction decoder\nleveraged for constructing trips for the vehicles. This policy model is\nequipped with two classes of state representations, fleet state and routing\nstate, providing the information needed for effective route construction in the\npresence of maximum working hours constraints. Experimental results using\nreal-world datasets from two major Canadian cities not only show that SED2AM\noutperforms the current state-of-the-art DRL-based and metaheuristic-based\nbaselines but also demonstrate its generalizability to solve larger-scale\nproblems.\n", "date": "2025-03-06", "categories": []} {"id": "2503.03338", "title": "Navigating Intelligence: A Survey of Google OR-Tools and Machine\n Learning for Global Path Planning in Autonomous Vehicles", "abstract": " We offer a new in-depth investigation of global path planning (GPP) for\nunmanned ground vehicles, an autonomous mining sampling robot named ROMIE. GPP\nis essential for ROMIE's optimal performance, which is translated into solving\nthe traveling salesman problem, a complex graph theory challenge that is\ncrucial for determining the most effective route to cover all sampling\nlocations in a mining field. This problem is central to enhancing ROMIE's\noperational efficiency and competitiveness against human labor by optimizing\ncost and time. The primary aim of this research is to advance GPP by\ndeveloping, evaluating, and improving a cost-efficient software and web\napplication. We delve into an extensive comparison and analysis of Google\noperations research (OR)-Tools optimization algorithms. Our study is driven by\nthe goal of applying and testing the limits of OR-Tools capabilities by\nintegrating Reinforcement Learning techniques for the first time. This enables\nus to compare these methods with OR-Tools, assessing their computational\neffectiveness and real-world application efficiency. Our analysis seeks to\nprovide insights into the effectiveness and practical application of each\ntechnique. Our findings indicate that Q-Learning stands out as the optimal\nstrategy, demonstrating superior efficiency by deviating only 1.2% on average\nfrom the optimal solutions across our datasets.\n", "date": "2025-03-05", "categories": []} {"id": "2503.03350", "title": "Leveraging Large Language Models to Develop Heuristics for Emerging\n Optimization Problems", "abstract": " Combinatorial optimization problems often rely on heuristic algorithms to\ngenerate efficient solutions. However, the manual design of heuristics is\nresource-intensive and constrained by the designer's expertise. Recent advances\nin artificial intelligence, particularly large language models (LLMs), have\ndemonstrated the potential to automate heuristic generation through\nevolutionary frameworks. Recent works focus only on well-known combinatorial\noptimization problems like the traveling salesman problem and online bin\npacking problem when designing constructive heuristics. This study investigates\nwhether LLMs can effectively generate heuristics for niche, not yet broadly\nresearched optimization problems, using the unit-load pre-marshalling problem\nas an example case. We propose the Contextual Evolution of Heuristics (CEoH)\nframework, an extension of the Evolution of Heuristics (EoH) framework, which\nincorporates problem-specific descriptions to enhance in-context learning\nduring heuristic generation. Through computational experiments, we evaluate\nCEoH and EoH and compare the results. Results indicate that CEoH enables\nsmaller LLMs to generate high-quality heuristics more consistently and even\noutperform larger models. Larger models demonstrate robust performance with or\nwithout contextualized prompts. The generated heuristics exhibit scalability to\ndiverse instance configurations.\n", "date": "2025-03-05", "categories": []} {"id": "2304.12146", "title": "Combining Monte Carlo Tree Search and Heuristic Search for Weighted\n Vertex Coloring", "abstract": " This work investigates the Monte Carlo Tree Search (MCTS) method combined\nwith dedicated heuristics for solving the Weighted Vertex Coloring Problem. In\naddition to the basic MCTS algorithm, we study several MCTS variants where the\nconventional random simulation is replaced by other simulation strategies\nincluding greedy and local search heuristics. We conduct experiments on\nwell-known benchmark instances to assess these combined MCTS variants. We\nprovide empirical evidence to shed light on the advantages and limits of each\nsimulation strategy. This is an extension of the work of Grelier and al.\npresented at EvoCOP2022.\n", "date": "2023-04-24", "categories": []} {"id": "2503.01931", "title": "Adversarial Generative Flow Network for Solving Vehicle Routing Problems", "abstract": " Recent research into solving vehicle routing problems (VRPs) has gained\nsignificant traction, particularly through the application of deep\n(reinforcement) learning for end-to-end solution construction. However, many\ncurrent construction-based neural solvers predominantly utilize Transformer\narchitectures, which can face scalability challenges and struggle to produce\ndiverse solutions. To address these limitations, we introduce a novel framework\nbeyond Transformer-based approaches, i.e., Adversarial Generative Flow Networks\n(AGFN). This framework integrates the generative flow network (GFlowNet)-a\nprobabilistic model inherently adept at generating diverse solutions\n(routes)-with a complementary model for discriminating (or evaluating) the\nsolutions. These models are trained alternately in an adversarial manner to\nimprove the overall solution quality, followed by a proposed hybrid decoding\nmethod to construct the solution. We apply the AGFN framework to solve the\ncapacitated vehicle routing problem (CVRP) and travelling salesman problem\n(TSP), and our experimental results demonstrate that AGFN surpasses the popular\nconstruction-based neural solvers, showcasing strong generalization\ncapabilities on synthetic and real-world benchmark instances.\n", "date": "2025-03-03", "categories": []} {"id": "2503.02180", "title": "Discrete Differential Evolution Particle Swarm Optimization Algorithm\n for Energy Saving Flexible Job Shop Scheduling Problem Considering Machine\n Multi States", "abstract": " As the continuous deepening of low-carbon emission reduction policies, the\nmanufacturing industries urgently need sensible energy-saving scheduling\nschemes to achieve the balance between improving production efficiency and\nreducing energy consumption. In energy-saving scheduling, reasonable machine\nstates-switching is a key point to achieve expected goals, i.e., whether the\nmachines need to switch speed between different operations, and whether the\nmachines need to add extra setup time between different jobs. Regarding this\nmatter, this work proposes a novel machine multi states-based energy saving\nflexible job scheduling problem (EFJSP-M), which simultaneously takes into\naccount machine multi speeds and setup time. To address the proposed EFJSP-M, a\nkind of discrete differential evolution particle swarm optimization algorithm\n(D-DEPSO) is designed. In specific, D-DEPSO includes a hybrid initialization\nstrategy to improve the initial population performance, an updating mechanism\nembedded with differential evolution operators to enhance population diversity,\nand a critical path variable neighborhood search strategy to expand the\nsolution space. At last, based on datasets DPs and MKs, the experiment results\ncompared with five state-of-the-art algorithms demonstrate the feasible of\nEFJSP-M and the superior of D-DEPSO.\n", "date": "2025-03-04", "categories": []} {"id": "2410.16282", "title": "Optimal Ground Station Selection for Low-Earth Orbiting Satellites", "abstract": " This paper presents a solution to the problem of optimal ground station\nselection for low-Earth orbiting (LEO) space missions that enables mission\noperators to precisely design their ground segment performance and costs. Space\nmission operators are increasingly turning to Ground-Station-as-a-Service\n(GSaaS) providers to supply the terrestrial communications segment to reduce\ncosts and increase network size. However, this approach leads to a new\nchallenge of selecting the optimal service providers and station locations for\na given mission. We consider the problem of ground station selection as an\noptimization problem and present a general solution framework that allows\nmission designers to set their overall optimization objective and constrain key\nmission performance variables such as total data downlink, total mission cost,\nrecurring operational cost, and maximum communications time-gap. We solve the\nproblem using integer programming (IP). To address computational scaling\nchallenges, we introduce a surrogate optimization approach where the optimal\nstation selection is determined based on solving the problem over a reduced\ntime domain. Two different IP formulations are evaluated using randomized\nselections of LEO satellites of varying constellation sizes. We consider the\nnetworks of the commercial GSaaS providers Atlas Space Operations, Amazon Web\nServices (AWS) Ground Station, Azure Orbital Ground Station, Kongsberg\nSatellite Services (KSAT), Leaf Space, and Viasat Real-Time Earth. We compare\nour results against standard operational practices of integrating with one or\ntwo primary ground station providers.\n", "date": "2024-10-04", "categories": []} {"id": "2503.00753", "title": "Rethinking Light Decoder-based Solvers for Vehicle Routing Problems", "abstract": " Light decoder-based solvers have gained popularity for solving vehicle\nrouting problems (VRPs) due to their efficiency and ease of integration with\nreinforcement learning algorithms. However, they often struggle with\ngeneralization to larger problem instances or different VRP variants. This\npaper revisits light decoder-based approaches, analyzing the implications of\ntheir reliance on static embeddings and the inherent challenges that arise.\nSpecifically, we demonstrate that in the light decoder paradigm, the encoder is\nimplicitly tasked with capturing information for all potential decision\nscenarios during solution construction within a single set of embeddings,\nresulting in high information density. Furthermore, our empirical analysis\nreveals that the overly simplistic decoder struggles to effectively utilize\nthis dense information, particularly as task complexity increases, which limits\ngeneralization to out-of-distribution (OOD) settings. Building on these\ninsights, we show that enhancing the decoder capacity, with a simple addition\nof identity mapping and a feed-forward layer, can considerably alleviate the\ngeneralization issue. Experimentally, our method significantly enhances the OOD\ngeneralization of light decoder-based approaches on large-scale instances and\ncomplex VRP variants, narrowing the gap with the heavy decoder paradigm. Our\ncode is available at: https://github.com/ziweileonhuang/reld-nco.\n", "date": "2025-03-02", "categories": []} {"id": "2503.01121", "title": "Hybrid Metaheuristic Vehicle Routing Problem for Security Dispatch\n Operations", "abstract": " This paper investigates the optimization of the Vehicle Routing Problem for\nSecurity Dispatch (VRPSD). VRPSD focuses on security and patrolling\napplications which involve challenging constraints including precise timing and\nstrict time windows. We propose three algorithms based on different\nmetaheuristics, which are Adaptive Large Neighborhood Search (ALNS), Tabu\nSearch (TS), and Threshold Accepting (TA). The first algorithm combines\nsingle-phase ALNS with TA, the second employs a multiphase ALNS with TA, and\nthe third integrates multiphase ALNS, TS, and TA. Experiments are conducted on\nan instance comprising 251 customer requests. The results demonstrate that the\nthird algorithm, the hybrid multiphase ALNS-TS-TA algorithm, delivers the best\nperformance. This approach simultaneously leverages the large-area search\ncapabilities of ALNS for exploration and effectively escapes local optima when\nthe multiphase ALNS is coupled with TS and TA. Furthermore, in our experiments,\nthe hybrid multiphase ALNS-TS-TA algorithm is the only one that shows potential\nfor improving results with increased computation time across all attempts.\n", "date": "2025-03-03", "categories": []} {"id": "2503.01129", "title": "Apollo-MILP: An Alternating Prediction-Correction Neural Solving\n Framework for Mixed-Integer Linear Programming", "abstract": " Leveraging machine learning (ML) to predict an initial solution for\nmixed-integer linear programming (MILP) has gained considerable popularity in\nrecent years. These methods predict a solution and fix a subset of variables to\nreduce the problem dimension. Then, they solve the reduced problem to obtain\nthe final solutions. However, directly fixing variable values can lead to\nlow-quality solutions or even infeasible reduced problems if the predicted\nsolution is not accurate enough. To address this challenge, we propose an\nAlternating prediction-correction neural solving framework (Apollo-MILP) that\ncan identify and select accurate and reliable predicted values to fix. In each\niteration, Apollo-MILP conducts a prediction step for the unfixed variables,\nfollowed by a correction step to obtain an improved solution (called reference\nsolution) through a trust-region search. By incorporating the predicted and\nreference solutions, we introduce a novel Uncertainty-based Error upper BOund\n(UEBO) to evaluate the uncertainty of the predicted values and fix those with\nhigh confidence. A notable feature of Apollo-MILP is the superior ability for\nproblem reduction while preserving optimality, leading to high-quality final\nsolutions. Experiments on commonly used benchmarks demonstrate that our\nproposed Apollo-MILP significantly outperforms other ML-based approaches in\nterms of solution quality, achieving over a 50% reduction in the solution gap.\n", "date": "2025-03-03", "categories": []} {"id": "2403.07041", "title": "Ant Colony Sampling with GFlowNets for Combinatorial Optimization", "abstract": " We present the Generative Flow Ant Colony Sampler (GFACS), a novel\nmeta-heuristic method that hierarchically combines amortized inference and\nparallel stochastic search. Our method first leverages Generative Flow Networks\n(GFlowNets) to amortize a \\emph{multi-modal} prior distribution over\ncombinatorial solution space that encompasses both high-reward and diversified\nsolutions. This prior is iteratively updated via parallel stochastic search in\nthe spirit of Ant Colony Optimization (ACO), leading to the posterior\ndistribution that generates near-optimal solutions. Extensive experiments\nacross seven combinatorial optimization problems demonstrate GFACS's promising\nperformances.\n", "date": "2024-03-11", "categories": []} {"id": "2412.15534", "title": "SORREL: Suboptimal-Demonstration-Guided Reinforcement Learning for\n Learning to Branch", "abstract": " Mixed Integer Linear Program (MILP) solvers are mostly built upon a\nBranch-and-Bound (B\\&B) algorithm, where the efficiency of traditional solvers\nheavily depends on hand-crafted heuristics for branching. The past few years\nhave witnessed the increasing popularity of data-driven approaches to\nautomatically learn these heuristics. However, the success of these methods is\nhighly dependent on the availability of high-quality demonstrations, which\nrequires either the development of near-optimal heuristics or a time-consuming\nsampling process. This paper averts this challenge by proposing\nSuboptimal-Demonstration-Guided Reinforcement Learning (SORREL) for learning to\nbranch. SORREL selectively learns from suboptimal demonstrations based on value\nestimation. It utilizes suboptimal demonstrations through both offline\nreinforcement learning on the demonstrations generated by suboptimal heuristics\nand self-imitation learning on past good experiences sampled by itself. Our\nexperiments demonstrate its advanced performance in both branching quality and\ntraining efficiency over previous methods for various MILPs.\n", "date": "2024-12-20", "categories": []} {"id": "2406.15079", "title": "GOAL: A Generalist Combinatorial Optimization Agent Learner", "abstract": " Machine Learning-based heuristics have recently shown impressive performance\nin solving a variety of hard combinatorial optimization problems (COPs).\nHowever, they generally rely on a separate neural model, specialized and\ntrained for each single problem. Any variation of a problem requires adjustment\nof its model and re-training from scratch. In this paper, we propose GOAL (for\nGeneralist combinatorial Optimization Agent Learner), a generalist model\ncapable of efficiently solving multiple COPs and which can be fine-tuned to\nsolve new COPs. GOAL consists of a single backbone plus light-weight\nproblem-specific adapters for input and output processing. The backbone is\nbased on a new form of mixed-attention blocks which allows to handle problems\ndefined on graphs with arbitrary combinations of node, edge and instance-level\nfeatures. Additionally, problems which involve heterogeneous types of nodes or\nedges are handled through a novel multi-type transformer architecture, where\nthe attention blocks are duplicated to attend the meaningful combinations of\ntypes while relying on the same shared parameters. We train GOAL on a set of\nrouting, scheduling and classic graph problems and show that it is only\nslightly inferior to the specialized baselines while being the first multi-task\nmodel that solves a wide range of COPs. Finally we showcase the strong transfer\nlearning capacity of GOAL by fine-tuning it on several new problems. Our code\nis available at https://github.com/naver/goal-co/.\n", "date": "2024-06-21", "categories": []} {"id": "2406.02858", "title": "TSPDiffuser: Diffusion Models as Learned Samplers for Traveling\n Salesperson Path Planning Problems", "abstract": " This paper presents TSPDiffuser, a novel data-driven path planner for\ntraveling salesperson path planning problems (TSPPPs) in environments rich with\nobstacles. Given a set of destinations within obstacle maps, our objective is\nto efficiently find the shortest possible collision-free path that visits all\nthe destinations. In TSPDiffuser, we train a diffusion model on a large\ncollection of TSPPP instances and their respective solutions to generate\nplausible paths for unseen problem instances. The model can then be employed as\na learned sampler to construct a roadmap that contains potential solutions with\na small number of nodes and edges. This approach enables efficient and accurate\nestimation of travel costs between destinations, effectively addressing the\nprimary computational challenge in solving TSPPPs. Experimental evaluations\nwith diverse synthetic and real-world indoor/outdoor environments demonstrate\nthe effectiveness of TSPDiffuser over existing methods in terms of the\ntrade-off between solution quality and computational time requirements.\n", "date": "2024-06-05", "categories": []} {"id": "2406.09657", "title": "Mitigating over-exploration in latent space optimization using LES", "abstract": " We develop Latent Exploration Score (LES) to mitigate over-exploration in\nLatent Space Optimization (LSO), a popular method for solving black-box\ndiscrete optimization problems. LSO utilizes continuous optimization within the\nlatent space of a Variational Autoencoder (VAE) and is known to be susceptible\nto over-exploration, which manifests in unrealistic solutions that reduce its\npracticality. LES leverages the trained decoder's approximation of the data\ndistribution, and can be employed with any VAE decoder - including pretrained\nones - without additional training, architectural changes or access to the\ntraining data. Our evaluation across five LSO benchmark tasks and twenty-two\nVAE models demonstrates that LES always enhances the quality of the solutions\nwhile maintaining high objective values, leading to improvements over existing\nsolutions in most cases. We believe that new avenues to LSO will be opened by\nLES' ability to identify out of distribution areas, differentiability, and\ncomputational tractability. Open source code for LES is available at\nhttps://github.com/OmerRonen/les.\n", "date": "2024-06-14", "categories": []} {"id": "2409.18582", "title": "Optimistic Games for Combinatorial Bayesian Optimization with\n Application to Protein Design", "abstract": " Bayesian optimization (BO) is a powerful framework to optimize black-box\nexpensive-to-evaluate functions via sequential interactions. In several\nimportant problems (e.g. drug discovery, circuit design, neural architecture\nsearch, etc.), though, such functions are defined over large\n$\\textit{combinatorial and unstructured}$ spaces. This makes existing BO\nalgorithms not feasible due to the intractable maximization of the acquisition\nfunction over these domains. To address this issue, we propose\n$\\textbf{GameOpt}$, a novel game-theoretical approach to combinatorial BO.\n$\\textbf{GameOpt}$ establishes a cooperative game between the different\noptimization variables, and selects points that are game $\\textit{equilibria}$\nof an upper confidence bound acquisition function. These are stable\nconfigurations from which no variable has an incentive to deviate$-$ analog to\nlocal optima in continuous domains. Crucially, this allows us to efficiently\nbreak down the complexity of the combinatorial domain into individual decision\nsets, making $\\textbf{GameOpt}$ scalable to large combinatorial spaces. We\ndemonstrate the application of $\\textbf{GameOpt}$ to the challenging\n$\\textit{protein design}$ problem and validate its performance on four\nreal-world protein datasets. Each protein can take up to $20^{X}$ possible\nconfigurations, where $X$ is the length of a protein, making standard BO\nmethods infeasible. Instead, our approach iteratively selects informative\nprotein configurations and very quickly discovers highly active protein\nvariants compared to other baselines.\n", "date": "2024-09-27", "categories": []} {"id": "2502.15777", "title": "TSS GAZ PTP: Towards Improving Gumbel AlphaZero with Two-stage Self-play\n for Multi-constrained Electric Vehicle Routing Problems", "abstract": " Recently, Gumbel AlphaZero~(GAZ) was proposed to solve classic combinatorial\noptimization problems such as TSP and JSSP by creating a carefully designed\ncompetition model~(consisting of a learning player and a competitor player),\nwhich leverages the idea of self-play. However, if the competitor is too strong\nor too weak, the effectiveness of self-play training can be reduced,\nparticularly in complex CO problems. To address this problem, we further\npropose a two-stage self-play strategy to improve the GAZ method~(named TSS GAZ\nPTP). In the first stage, the learning player uses the enhanced policy network\nbased on the Gumbel Monte Carlo Tree Search~(MCTS), and the competitor uses the\nhistorical best trained policy network~(acts as a greedy player). In the second\nstage, we employ Gumbel MCTS for both players, which makes the competition\nfiercer so that both players can continuously learn smarter trajectories. We\nfirst investigate the performance of our proposed TSS GAZ PTP method on TSP\nsince it is also used as a test problem by the original GAZ. The results show\nthe superior performance of TSS GAZ PTP. Then we extend TSS GAZ PTP to deal\nwith multi-constrained Electric Vehicle Routing Problems~(EVRP), which is a\nrecently well-known real application research topic and remains challenging as\na complex CO problem. Impressively, the experimental results show that the TSS\nGAZ PTP outperforms the state-of-the-art Deep Reinforcement Learning methods in\nall types of instances tested and outperforms the optimization solver in tested\nlarge-scale instances, indicating the importance and promising of employing\nmore dynamic self-play strategies for complex CO problems.\n", "date": "2025-02-17", "categories": []} {"id": "2502.15791", "title": "Learning-Guided Rolling Horizon Optimization for Long-Horizon Flexible\n Job-Shop Scheduling", "abstract": " Long-horizon combinatorial optimization problems (COPs), such as the Flexible\nJob-Shop Scheduling Problem (FJSP), often involve complex, interdependent\ndecisions over extended time frames, posing significant challenges for existing\nsolvers. While Rolling Horizon Optimization (RHO) addresses this by decomposing\nproblems into overlapping shorter-horizon subproblems, such overlap often\ninvolves redundant computations. In this paper, we present L-RHO, the first\nlearning-guided RHO framework for COPs. L-RHO employs a neural network to\nintelligently fix variables that in hindsight did not need to be re-optimized,\nresulting in smaller and thus easier-to-solve subproblems. For FJSP, this means\nidentifying operations with unchanged machine assignments between consecutive\nsubproblems. Applied to FJSP, L-RHO accelerates RHO by up to 54% while\nsignificantly improving solution quality, outperforming other heuristic and\nlearning-based baselines. We also provide in-depth discussions and verify the\ndesirable adaptability and generalization of L-RHO across numerous FJSP\nvariates, distributions, online scenarios and benchmark instances. Moreover, we\nprovide a theoretical analysis to elucidate the conditions under which learning\nis beneficial.\n", "date": "2025-02-18", "categories": []} {"id": "2502.16170", "title": "Destroy and Repair Using Hyper Graphs for Routing", "abstract": " Recent advancements in Neural Combinatorial Optimization (NCO) have shown\npromise in solving routing problems like the Traveling Salesman Problem (TSP)\nand Capacitated Vehicle Routing Problem (CVRP) without handcrafted designs.\nResearch in this domain has explored two primary categories of methods:\niterative and non-iterative. While non-iterative methods struggle to generate\nnear-optimal solutions directly, iterative methods simplify the task by\nlearning local search steps. However, existing iterative methods are often\nlimited by restricted neighborhood searches, leading to suboptimal results. To\naddress this limitation, we propose a novel approach that extends the search to\nlarger neighborhoods by learning a destroy-and-repair strategy. Specifically,\nwe introduce a Destroy-and-Repair framework based on Hyper-Graphs (DRHG). This\nframework reduces consecutive intact edges to hyper-edges, allowing the model\nto pay more attention to the destroyed part and decrease the complexity of\nencoding all nodes. Experiments demonstrate that DRHG achieves stateof-the-art\nperformance on TSP with up to 10,000 nodes and shows strong generalization to\nreal-world TSPLib and CVRPLib problems.\n", "date": "2025-02-22", "categories": []} {"id": "2502.14102", "title": "Explainable Distributed Constraint Optimization Problems", "abstract": " The Distributed Constraint Optimization Problem (DCOP) formulation is a\npowerful tool to model cooperative multi-agent problems that need to be solved\ndistributively. A core assumption of existing approaches is that DCOP solutions\ncan be easily understood, accepted, and adopted, which may not hold, as\nevidenced by the large body of literature on Explainable AI. In this paper, we\npropose the Explainable DCOP (X-DCOP) model, which extends a DCOP to include\nits solution and a contrastive query for that solution. We formally define some\nkey properties that contrastive explanations must satisfy for them to be\nconsidered as valid solutions to X-DCOPs as well as theoretical results on the\nexistence of such valid explanations. To solve X-DCOPs, we propose a\ndistributed framework as well as several optimizations and suboptimal variants\nto find valid explanations. We also include a human user study that showed that\nusers, not surprisingly, prefer shorter explanations over longer ones. Our\nempirical evaluations showed that our approach can scale to large problems, and\nthe different variants provide different options for trading off explanation\nlengths for smaller runtimes. Thus, our model and algorithmic contributions\nextend the state of the art by reducing the barrier for users to understand\nDCOP solutions, facilitating their adoption in more real-world applications.\n", "date": "2025-02-19", "categories": []} {"id": "2502.13266", "title": "A Machine Learning Approach That Beats Large Rubik's Cubes", "abstract": " The paper proposes a novel machine learning-based approach to the pathfinding\nproblem on extremely large graphs. This method leverages diffusion distance\nestimation via a neural network and uses beam search for pathfinding. We\ndemonstrate its efficiency by finding solutions for 4x4x4 and 5x5x5 Rubik's\ncubes with unprecedentedly short solution lengths, outperforming all available\nsolvers and introducing the first machine learning solver beyond the 3x3x3\ncase. In particular, it surpasses every single case of the combined best\nresults in the Kaggle Santa 2023 challenge, which involved over 1,000 teams.\nFor the 3x3x3 Rubik's cube, our approach achieves an optimality rate exceeding\n98%, matching the performance of task-specific solvers and significantly\noutperforming prior solutions such as DeepCubeA (60.3%) and EfficientCube\n(69.6%). Additionally, our solution is more than 26 times faster in solving\n3x3x3 Rubik's cubes while requiring up to 18.5 times less model training time\nthan the most efficient state-of-the-art competitor.\n", "date": "2025-02-18", "categories": []} {"id": "2502.13769", "title": "A consensus set for the aggregation of partial rankings: the case of the\n Optimal Set of Bucket Orders Problem", "abstract": " In rank aggregation problems (RAP), the solution is usually a consensus\nranking that generalizes a set of input orderings. There are different variants\nthat differ not only in terms of the type of rankings that are used as input\nand output, but also in terms of the objective function employed to evaluate\nthe quality of the desired output ranking. In contrast, in some machine\nlearning tasks (e.g. subgroup discovery) or multimodal optimization tasks,\nattention is devoted to obtaining several models/results to account for the\ndiversity in the input data or across the search landscape. Thus, in this paper\nwe propose to provide, as the solution to an RAP, a set of rankings to better\nexplain the preferences expressed in the input orderings. We exemplify our\nproposal through the Optimal Bucket Order Problem (OBOP), an RAP which consists\nin finding a single consensus ranking (with ties) that generalizes a set of\ninput rankings codified as a precedence matrix. To address this, we introduce\nthe Optimal Set of Bucket Orders Problem (OSBOP), a generalization of the OBOP\nthat aims to produce not a single ranking as output but a set of consensus\nrankings. Experimental results are presented to illustrate this proposal,\nshowing how, by providing a set of consensus rankings, the fitness of the\nsolution significantly improves with respect to the one of the original OBOP,\nwithout losing comprehensibility.\n", "date": "2025-02-19", "categories": []} {"id": "2305.08353", "title": "Fast and Efficient Matching Algorithm with Deadline Instances", "abstract": " The online weighted matching problem is a fundamental problem in machine\nlearning due to its numerous applications. Despite many efforts in this area,\nexisting algorithms are either too slow or don't take $\\mathrm{deadline}$ (the\nlongest time a node can be matched) into account. In this paper, we introduce a\nmarket model with $\\mathrm{deadline}$ first. Next, we present our two optimized\nalgorithms (\\textsc{FastGreedy} and \\textsc{FastPostponedGreedy}) and offer\ntheoretical proof of the time complexity and correctness of our algorithms. In\n\\textsc{FastGreedy} algorithm, we have already known if a node is a buyer or a\nseller. But in \\textsc{FastPostponedGreedy} algorithm, the status of each node\nis unknown at first. Then, we generalize a sketching matrix to run the original\nand our algorithms on both real data sets and synthetic data sets. Let\n$\\epsilon \\in (0,0.1)$ denote the relative error of the real weight of each\nedge. The competitive ratio of original \\textsc{Greedy} and\n\\textsc{PostponedGreedy} is $\\frac{1}{2}$ and $\\frac{1}{4}$ respectively. Based\non these two original algorithms, we proposed \\textsc{FastGreedy} and\n\\textsc{FastPostponedGreedy} algorithms and the competitive ratio of them is\n$\\frac{1 - \\epsilon}{2}$ and $\\frac{1 - \\epsilon}{4}$ respectively. At the same\ntime, our algorithms run faster than the original two algorithms. Given $n$\nnodes in $\\mathbb{R} ^ d$, we decrease the time complexity from $O(nd)$ to\n$\\widetilde{O}(\\epsilon^{-2} \\cdot (n + d))$, where for any function $f$, we\nuse $\\widetilde{O}(f)$ to denote $f \\cdot \\mathrm{poly}(\\log f)$.\n", "date": "2023-05-15", "categories": []} {"id": "2502.12993", "title": "Approximate Tree Completion and Learning-Augmented Algorithms for Metric\n Minimum Spanning Trees", "abstract": " Finding a minimum spanning tree (MST) for $n$ points in an arbitrary metric\nspace is a fundamental primitive for hierarchical clustering and many other ML\ntasks, but this takes $\\Omega(n^2)$ time to even approximate. We introduce a\nframework for metric MSTs that first (1) finds a forest of disconnected\ncomponents using practical heuristics, and then (2) finds a small weight set of\nedges to connect disjoint components of the forest into a spanning tree. We\nprove that optimally solving the second step still takes $\\Omega(n^2)$ time,\nbut we provide a subquadratic 2.62-approximation algorithm. In the spirit of\nlearning-augmented algorithms, we then show that if the forest found in step\n(1) overlaps with an optimal MST, we can approximate the original MST problem\nin subquadratic time, where the approximation factor depends on a measure of\noverlap. In practice, we find nearly optimal spanning trees for a wide range of\nmetrics, while being orders of magnitude faster than exact algorithms.\n", "date": "2025-02-18", "categories": []} {"id": "2502.13000", "title": "Edge-Colored Clustering in Hypergraphs: Beyond Minimizing Unsatisfied\n Edges", "abstract": " We consider a framework for clustering edge-colored hypergraphs, where the\ngoal is to cluster (equivalently, to color) objects based on the primary type\nof multiway interactions they participate in. One well-studied objective is to\ncolor nodes to minimize the number of unsatisfied hyperedges -- those\ncontaining one or more nodes whose color does not match the hyperedge color. We\nmotivate and present advances for several directions that extend beyond this\nminimization problem. We first provide new algorithms for maximizing satisfied\nedges, which is the same at optimality but is much more challenging to\napproximate, with all prior work restricted to graphs. We develop the first\napproximation algorithm for hypergraphs, and then refine it to improve the\nbest-known approximation factor for graphs. We then introduce new objective\nfunctions that incorporate notions of balance and fairness, and provide new\nhardness results, approximations, and fixed-parameter tractability results.\n", "date": "2025-02-18", "categories": []} {"id": "2403.10332", "title": "GreedyML: A Parallel Algorithm for Maximizing Constrained Submodular\n Functions", "abstract": " We describe a parallel approximation algorithm for maximizing monotone\nsubmodular functions subject to hereditary constraints on distributed memory\nmultiprocessors. Our work is motivated by the need to solve submodular\noptimization problems on massive data sets, for practical contexts such as data\nsummarization, machine learning, and graph sparsification.\n Our work builds on the randomized distributed RandGreedi algorithm, proposed\nby Barbosa, Ene, Nguyen, and Ward (2015). This algorithm computes a distributed\nsolution by randomly partitioning the data among all the processors and then\nemploying \\emph{a single} accumulation step in which all processors send their\npartial solutions to one processor. However, for large problems, the\naccumulation step exceeds the memory available on a processor, and the\nprocessor that performs the accumulation becomes a computational bottleneck.\n Hence we propose a generalization of the RandGreedi algorithm that employs\nmultiple accumulation steps to reduce the memory required. We analyze the\napproximation ratio and the time complexity of the algorithm (in the BSP\nmodel). We evaluate the new GreedyML algorithm on three classes of problems,\nand report results from large-scale data sets with millions of elements. The\nresults show that the GreedyML algorithm can solve problems where the\nsequential Greedy and distributed RandGreedi algorithms fail due to memory\nconstraints. For certain computationally intensive problems, the GreedyML\nalgorithm is faster than the RandGreedi algorithm. The observed approximation\nquality of the solutions computed by the GreedyML algorithm closely matches\nthose obtained by the RandGreedi algorithm on these problems.\n", "date": "2024-03-15", "categories": []} {"id": "2502.11134", "title": "Solving Online Resource-Constrained Scheduling for Follow-Up Observation\n in Astronomy: a Reinforcement Learning Approach", "abstract": " In the astronomical observation field, determining the allocation of\nobservation resources of the telescope array and planning follow-up\nobservations for targets of opportunity (ToOs) are indispensable components of\nastronomical scientific discovery. This problem is computationally challenging,\ngiven the online observation setting and the abundance of time-varying factors\nthat can affect whether an observation can be conducted. This paper presents\nROARS, a reinforcement learning approach for online astronomical\nresource-constrained scheduling. To capture the structure of the astronomical\nobservation scheduling, we depict every schedule using a directed acyclic graph\n(DAG), illustrating the dependency of timing between different observation\ntasks within the schedule. Deep reinforcement learning is used to learn a\npolicy that can improve the feasible solution by iteratively local rewriting\nuntil convergence. It can solve the challenge of obtaining a complete solution\ndirectly from scratch in astronomical observation scenarios, due to the high\ncomputational complexity resulting from numerous spatial and temporal\nconstraints. A simulation environment is developed based on real-world\nscenarios for experiments, to evaluate the effectiveness of our proposed\nscheduling approach. The experimental results show that ROARS surpasses 5\npopular heuristics, adapts to various observation scenarios and learns\neffective strategies with hindsight.\n", "date": "2025-02-16", "categories": []} {"id": "2502.11715", "title": "Proactive Depot Discovery: A Generative Framework for Flexible\n Location-Routing", "abstract": " The Location-Routing Problem (LRP), which combines the challenges of facility\n(depot) locating and vehicle route planning, is critically constrained by the\nreliance on predefined depot candidates, limiting the solution space and\npotentially leading to suboptimal outcomes. Previous research on LRP without\npredefined depots is scant and predominantly relies on heuristic algorithms\nthat iteratively attempt depot placements across a planar area. Such approaches\nlack the ability to proactively generate depot locations that meet specific\ngeographic requirements, revealing a notable gap in current research landscape.\nTo bridge this gap, we propose a data-driven generative DRL framework, designed\nto proactively generate depots for LRP without predefined depot candidates,\nsolely based on customer requests data which include geographic and demand\ninformation. It can operate in two distinct modes: direct generation of exact\ndepot locations, and the creation of a multivariate Gaussian distribution for\nflexible depots sampling. By extracting depots' geographic pattern from\ncustomer requests data, our approach can dynamically respond to logistical\nneeds, identifying high-quality depot locations that further reduce total\nrouting costs compared to traditional methods. Extensive experiments\ndemonstrate that, for a same group of customer requests, compared with those\ndepots identified through random attempts, our framework can proactively\ngenerate depots that lead to superior solution routes with lower routing cost.\nThe implications of our framework potentially extend into real-world\napplications, particularly in emergency medical rescue and disaster relief\nlogistics, where rapid establishment and adjustment of depot locations are\nparamount, showcasing its potential in addressing LRP for dynamic and\nunpredictable environments.\n", "date": "2025-02-17", "categories": []} {"id": "2502.10226", "title": "A Multiagent Path Search Algorithm for Large-Scale Coalition Structure\n Generation", "abstract": " Coalition structure generation (CSG), i.e. the problem of optimally\npartitioning a set of agents into coalitions to maximize social welfare, is a\nfundamental computational problem in multiagent systems. This problem is\nimportant for many applications where small run times are necessary, including\ntransportation and disaster response. In this paper, we develop SALDAE, a\nmultiagent path finding algorithm for CSG that operates on a graph of coalition\nstructures. Our algorithm utilizes a variety of heuristics and strategies to\nperform the search and guide it. It is an anytime algorithm that can handle\nlarge problems with hundreds and thousands of agents. We show empirically on\nnine standard value distributions, including disaster response and electric\nvehicle allocation benchmarks, that our algorithm enables a rapid finding of\nhigh-quality solutions and compares favorably with other state-of-the-art\nmethods.\n", "date": "2025-02-14", "categories": []} {"id": "2502.10233", "title": "Learning to Solve the Min-Max Mixed-Shelves Picker-Routing Problem via\n Hierarchical and Parallel Decoding", "abstract": " The Mixed-Shelves Picker Routing Problem (MSPRP) is a fundamental challenge\nin warehouse logistics, where pickers must navigate a mixed-shelves environment\nto retrieve SKUs efficiently. Traditional heuristics and optimization-based\napproaches struggle with scalability, while recent machine learning methods\noften rely on sequential decision-making, leading to high solution latency and\nsuboptimal agent coordination. In this work, we propose a novel hierarchical\nand parallel decoding approach for solving the min-max variant of the MSPRP via\nmulti-agent reinforcement learning. While our approach generates a joint\ndistribution over agent actions, allowing for fast decoding and effective\npicker coordination, our method introduces a sequential action selection to\navoid conflicts in the multi-dimensional action space. Experiments show\nstate-of-the-art performance in both solution quality and inference speed,\nparticularly for large-scale and out-of-distribution instances. Our code is\npublicly available at http://github.com/LTluttmann/marl4msprp.\n", "date": "2025-02-14", "categories": []} {"id": "2502.10308", "title": "LLM-Powered Preference Elicitation in Combinatorial Assignment", "abstract": " We study the potential of large language models (LLMs) as proxies for humans\nto simplify preference elicitation (PE) in combinatorial assignment. While\ntraditional PE methods rely on iterative queries to capture preferences, LLMs\noffer a one-shot alternative with reduced human effort. We propose a framework\nfor LLM proxies that can work in tandem with SOTA ML-powered preference\nelicitation schemes. Our framework handles the novel challenges introduced by\nLLMs, such as response variability and increased computational costs. We\nexperimentally evaluate the efficiency of LLM proxies against human queries in\nthe well-studied course allocation domain, and we investigate the model\ncapabilities required for success. We find that our approach improves\nallocative efficiency by up to 20%, and these results are robust across\ndifferent LLMs and to differences in quality and accuracy of reporting.\n", "date": "2025-02-14", "categories": []} {"id": "2502.08684", "title": "Self-Evaluation for Job-Shop Scheduling", "abstract": " Combinatorial optimization problems, such as scheduling and route planning,\nare crucial in various industries but are computationally intractable due to\ntheir NP-hard nature. Neural Combinatorial Optimization methods leverage\nmachine learning to address these challenges but often depend on sequential\ndecision-making, which is prone to error accumulation as small mistakes\npropagate throughout the process. Inspired by self-evaluation techniques in\nLarge Language Models, we propose a novel framework that generates and\nevaluates subsets of assignments, moving beyond traditional stepwise\napproaches. Applied to the Job-Shop Scheduling Problem, our method integrates a\nheterogeneous graph neural network with a Transformer to build a policy model\nand a self-evaluation function. Experimental validation on challenging,\nwell-known benchmarks demonstrates the effectiveness of our approach,\nsurpassing state-of-the-art methods.\n", "date": "2025-02-12", "categories": []} {"id": "2405.18272", "title": "Metaheuristics and Large Language Models Join Forces: Toward an\n Integrated Optimization Approach", "abstract": " Since the rise of Large Language Models (LLMs) a couple of years ago,\nresearchers in metaheuristics (MHs) have wondered how to use their power in a\nbeneficial way within their algorithms. This paper introduces a novel approach\nthat leverages LLMs as pattern recognition tools to improve MHs. The resulting\nhybrid method, tested in the context of a social network-based combinatorial\noptimization problem, outperforms existing state-of-the-art approaches that\ncombine machine learning with MHs regarding the obtained solution quality. By\ncarefully designing prompts, we demonstrate that the output obtained from LLMs\ncan be used as problem knowledge, leading to improved results. Lastly, we\nacknowledge LLMs' potential drawbacks and limitations and consider it essential\nto examine them to advance this type of research further. Our method can be\nreproduced using a tool available at: https://github.com/camilochs/optipattern.\n", "date": "2024-05-28", "categories": []} {"id": "2411.00003", "title": "Unsupervised Training of Diffusion Models for Feasible Solution\n Generation in Neural Combinatorial Optimization", "abstract": " Recent advancements in neural combinatorial optimization (NCO) methods have\nshown promising results in generating near-optimal solutions without the need\nfor expert-crafted heuristics. However, high performance of these approaches\noften rely on problem-specific human-expertise-based search after generating\ncandidate solutions, limiting their applicability to commonly solved CO\nproblems such as Traveling Salesman Problem (TSP). In this paper, we present\nIC/DC, an unsupervised CO framework that directly trains a diffusion model from\nscratch. We train our model in a self-supervised way to minimize the cost of\nthe solution while adhering to the problem-specific constraints. IC/DC is\nspecialized in addressing CO problems involving two distinct sets of items, and\nit does not need problem-specific search processes to generate valid solutions.\nIC/DC employs a novel architecture capable of capturing the intricate\nrelationships between items, and thereby enabling effective optimization in\nchallenging CO scenarios. IC/DC achieves state-of-the-art performance relative\nto existing NCO methods on the Parallel Machine Scheduling Problem (PMSP) and\nAsymmetric Traveling Salesman Problem (ATSP).\n", "date": "2024-10-15", "categories": []} {"id": "2502.07819", "title": "Enhancing kidney transplantation through multi-agent kidney exchange\n programs: A comprehensive review and optimization models", "abstract": " This paper presents a comprehensive review of the last two decades of\nresearch on Kidney Exchange Programs (KEPs), systematically categorizing and\nclassifying key contributions to provide readers with a structured\nunderstanding of advancements in the field. The review highlights the evolution\nof KEP methodologies and lays the foundation for our contribution. We propose\nthree mathematical models aimed at improving both the quantity and quality of\nkidney transplants. Model 1 maximizes the number of transplants by focusing on\ncompatibility based on blood type and PRA, without additional constraints.\nModel 2 introduces a minimum Human Leukocyte Antigen (HLA) compatibility\nthreshold to enhance transplant quality, though this leads to fewer matches.\nModel 3 extends the problem to a Multi-Agent Kidney Exchange Program (MKEP),\npooling incompatible donor-recipient pairs across multiple agents, resulting in\na higher number of successful transplants while ensuring fairness across\nagents. Sensitivity analyses demonstrate trade-offs between transplant quantity\nand quality, with Model 3 striking the optimal balance by leveraging\nmulti-agent collaboration to improve both the number and quality of\ntransplants. These findings underscore the potential benefits of more\nintegrated kidney exchange systems.\n", "date": "2025-02-10", "categories": []} {"id": "2502.08298", "title": "Improving Existing Optimization Algorithms with LLMs", "abstract": " The integration of Large Language Models (LLMs) into optimization has created\na powerful synergy, opening exciting research opportunities. This paper\ninvestigates how LLMs can enhance existing optimization algorithms. Using their\npre-trained knowledge, we demonstrate their ability to propose innovative\nheuristic variations and implementation strategies. To evaluate this, we\napplied a non-trivial optimization algorithm, Construct, Merge, Solve and Adapt\n(CMSA) -- a hybrid metaheuristic for combinatorial optimization problems that\nincorporates a heuristic in the solution construction phase. Our results show\nthat an alternative heuristic proposed by GPT-4o outperforms the\nexpert-designed heuristic of CMSA, with the performance gap widening on larger\nand denser graphs. Project URL: https://imp-opt-algo-llms.surge.sh/\n", "date": "2025-02-12", "categories": []} {"id": "2502.08340", "title": "Hierarchical Learning-based Graph Partition for Large-scale Vehicle\n Routing Problems", "abstract": " Neural solvers based on the divide-and-conquer approach for Vehicle Routing\nProblems (VRPs) in general, and capacitated VRP (CVRP) in particular,\nintegrates the global partition of an instance with local constructions for\neach subproblem to enhance generalization. However, during the global partition\nphase, misclusterings within subgraphs have a tendency to progressively\ncompound throughout the multi-step decoding process of the learning-based\npartition policy. This suboptimal behavior in the global partition phase, in\nturn, may lead to a dramatic deterioration in the performance of the overall\ndecomposition-based system, despite using optimal local constructions. To\naddress these challenges, we propose a versatile Hierarchical Learning-based\nGraph Partition (HLGP) framework, which is tailored to benefit the partition of\nCVRP instances by synergistically integrating global and local partition\npolicies. Specifically, the global partition policy is tasked with creating the\ncoarse multi-way partition to generate the sequence of simpler two-way\npartition subtasks. These subtasks mark the initiation of the subsequent K\nlocal partition levels. At each local partition level, subtasks exclusive for\nthis level are assigned to the local partition policy which benefits from the\ninsensitive local topological features to incrementally alleviate the\ncompounded errors. This framework is versatile in the sense that it optimizes\nthe involved partition policies towards a unified objective harmoniously\ncompatible with both reinforcement learning (RL) and supervised learning (SL).\n(*Due to the notification of arXiv \"The Abstract field cannot be longer than\n1,920 characters\", the appeared Abstract is shortened. For the full Abstract,\nplease download the Article.)\n", "date": "2025-02-12", "categories": []} {"id": "2311.08170", "title": "Neural Lattice Reduction: A Self-Supervised Geometric Deep Learning\n Approach", "abstract": " Lattice reduction is a combinatorial optimization problem aimed at finding\nthe most orthogonal basis in a given lattice. The Lenstra-Lenstra-Lov\\'asz\n(LLL) algorithm is the best algorithm in the literature for solving this\nproblem. In light of recent research on algorithm discovery, in this work, we\nwould like to answer this question: is it possible to parametrize the algorithm\nspace for lattice reduction problem with neural networks and find an algorithm\nwithout supervised data? Our strategy is to use equivariant and invariant\nparametrizations and train in a self-supervised way. We design a deep neural\nmodel outputting factorized unimodular matrices and train it in a\nself-supervised manner by penalizing non-orthogonal lattice bases. We\nincorporate the symmetries of lattice reduction into the model by making it\ninvariant to isometries and scaling of the ambient space and equivariant with\nrespect to the hyperocrahedral group permuting and flipping the lattice basis\nelements. We show that this approach yields an algorithm with comparable\ncomplexity and performance to the LLL algorithm on a set of benchmarks.\nAdditionally, motivated by certain applications for wireless communication, we\nextend our method to a convolutional architecture which performs joint\nreduction of spatially-correlated lattices arranged in a grid, thereby\namortizing its cost over multiple lattices.\n", "date": "2023-11-14", "categories": []} {"id": "2502.05537", "title": "Sequential Stochastic Combinatorial Optimization Using Hierarchal\n Reinforcement Learning", "abstract": " Reinforcement learning (RL) has emerged as a promising tool for combinatorial\noptimization (CO) problems due to its ability to learn fast, effective, and\ngeneralizable solutions. Nonetheless, existing works mostly focus on one-shot\ndeterministic CO, while sequential stochastic CO (SSCO) has rarely been studied\ndespite its broad applications such as adaptive influence maximization (IM) and\ninfectious disease intervention. In this paper, we study the SSCO problem where\nwe first decide the budget (e.g., number of seed nodes in adaptive IM)\nallocation for all time steps, and then select a set of nodes for each time\nstep. The few existing studies on SSCO simplify the problems by assuming a\nuniformly distributed budget allocation over the time horizon, yielding\nsuboptimal solutions. We propose a generic hierarchical RL (HRL) framework\ncalled wake-sleep option (WS-option), a two-layer option-based framework that\nsimultaneously decides adaptive budget allocation on the higher layer and node\nselection on the lower layer. WS-option starts with a coherent formulation of\nthe two-layer Markov decision processes (MDPs), capturing the interdependencies\nbetween the two layers of decisions. Building on this, WS-option employs\nseveral innovative designs to balance the model's training stability and\ncomputational efficiency, preventing the vicious cyclic interference issue\nbetween the two layers. Empirical results show that WS-option exhibits\nsignificantly improved effectiveness and generalizability compared to\ntraditional methods. Moreover, the learned model can be generalized to larger\ngraphs, which significantly reduces the overhead of computational resources.\n", "date": "2025-02-08", "categories": []} {"id": "2502.04928", "title": "Generative-enhanced optimization for knapsack problems: an\n industry-relevant study", "abstract": " Optimization is a crucial task in various industries such as logistics,\naviation, manufacturing, chemical, pharmaceutical, and insurance, where finding\nthe best solution to a problem can result in significant cost savings and\nincreased efficiency. Tensor networks (TNs) have gained prominence in recent\nyears in modeling classical systems with quantum-inspired approaches. More\nrecently, TN generative-enhanced optimization (TN-GEO) has been proposed as a\nstrategy which uses generative modeling to efficiently sample valid solutions\nwith respect to certain constraints of optimization problems. Moreover, it has\nbeen shown that symmetric TNs (STNs) can encode certain constraints of\noptimization problems, thus aiding in their solution process. In this work, we\ninvestigate the applicability of TN- and STN-GEO to an industry relevant\nproblem class, a multi-knapsack problem, in which each object must be assigned\nto an available knapsack. We detail a prescription for practitioners to use the\nTN-and STN-GEO methodology and study its scaling behavior and dependence on its\nhyper-parameters. We benchmark 60 different problem instances and find that\nTN-GEO and STN-GEO produce results of similar quality to simulated annealing.\n", "date": "2025-02-07", "categories": []} {"id": "2405.05202", "title": "Discretely Beyond $1/e$: Guided Combinatorial Algorithms for Submodular\n Maximization", "abstract": " For constrained, not necessarily monotone submodular maximization, all known\napproximation algorithms with ratio greater than $1/e$ require continuous\nideas, such as queries to the multilinear extension of a submodular function\nand its gradient, which are typically expensive to simulate with the original\nset function. For combinatorial algorithms, the best known approximation ratios\nfor both size and matroid constraint are obtained by a simple randomized greedy\nalgorithm of Buchbinder et al. [9]: $1/e \\approx 0.367$ for size constraint and\n$0.281$ for the matroid constraint in $\\mathcal O (kn)$ queries, where $k$ is\nthe rank of the matroid. In this work, we develop the first combinatorial\nalgorithms to break the $1/e$ barrier: we obtain approximation ratio of $0.385$\nin $\\mathcal O (kn)$ queries to the submodular set function for size\nconstraint, and $0.305$ for a general matroid constraint. These are achieved by\nguiding the randomized greedy algorithm with a fast local search algorithm.\nFurther, we develop deterministic versions of these algorithms, maintaining the\nsame ratio and asymptotic time complexity. Finally, we develop a deterministic,\nnearly linear time algorithm with ratio $0.377$.\n", "date": "2024-05-08", "categories": []} {"id": "2502.02941", "title": "Fast T2T: Optimization Consistency Speeds Up Diffusion-Based\n Training-to-Testing Solving for Combinatorial Optimization", "abstract": " Diffusion models have recently advanced Combinatorial Optimization (CO) as a\npowerful backbone for neural solvers. However, their iterative sampling process\nrequiring denoising across multiple noise levels incurs substantial overhead.\nWe propose to learn direct mappings from different noise levels to the optimal\nsolution for a given instance, facilitating high-quality generation with\nminimal shots. This is achieved through an optimization consistency training\nprotocol, which, for a given instance, minimizes the difference among samples\noriginating from varying generative trajectories and time steps relative to the\noptimal solution. The proposed model enables fast single-step solution\ngeneration while retaining the option of multi-step sampling to trade for\nsampling quality, which offers a more effective and efficient alternative\nbackbone for neural solvers. In addition, within the training-to-testing (T2T)\nframework, to bridge the gap between training on historical instances and\nsolving new instances, we introduce a novel consistency-based gradient search\nscheme during the test stage, enabling more effective exploration of the\nsolution space learned during training. It is achieved by updating the latent\nsolution probabilities under objective gradient guidance during the alternation\nof noise injection and denoising steps. We refer to this model as Fast T2T.\nExtensive experiments on two popular tasks, the Traveling Salesman Problem\n(TSP) and Maximal Independent Set (MIS), demonstrate the superiority of Fast\nT2T regarding both solution quality and efficiency, even outperforming LKH\ngiven limited time budgets. Notably, Fast T2T with merely one-step generation\nand one-step gradient search can mostly outperform the SOTA diffusion-based\ncounterparts that require hundreds of steps, while achieving tens of times\nspeedup.\n", "date": "2025-02-05", "categories": []} {"id": "2409.02697", "title": "Decision Transformer for Enhancing Neural Local Search on the Job Shop\n Scheduling Problem", "abstract": " The job shop scheduling problem (JSSP) and its solution algorithms have been\nof enduring interest in both academia and industry for decades. In recent\nyears, machine learning (ML) is playing an increasingly important role in\nadvancing existing and building new heuristic solutions for the JSSP, aiming to\nfind better solutions in shorter computation times. In this paper we build on\ntop of a state-of-the-art deep reinforcement learning (DRL) agent, called\nNeural Local Search (NLS), which can efficiently and effectively control a\nlarge local neighborhood search on the JSSP. In particular, we develop a method\nfor training the decision transformer (DT) algorithm on search trajectories\ntaken by a trained NLS agent to further improve upon the learned\ndecision-making sequences. Our experiments show that the DT successfully learns\nlocal search strategies that are different and, in many cases, more effective\nthan those of the NLS agent itself. In terms of the tradeoff between solution\nquality and acceptable computational time needed for the search, the DT is\nparticularly superior in application scenarios where longer computational times\nare acceptable. In this case, it makes up for the longer inference times\nrequired per search step, which are caused by the larger neural network\narchitecture, through better quality decisions per step. Thereby, the DT\nachieves state-of-the-art results for solving the JSSP with ML-enhanced search.\n", "date": "2024-09-04", "categories": []} {"id": "2409.16867", "title": "Multi-objective Evolution of Heuristic Using Large Language Model", "abstract": " Heuristics are commonly used to tackle various search and optimization\nproblems. Design heuristics usually require tedious manual crafting with domain\nknowledge. Recent works have incorporated Large Language Models (LLMs) into\nautomatic heuristic search, leveraging their powerful language and coding\ncapacity. However, existing research focuses on the optimal performance on the\ntarget problem as the sole objective, neglecting other criteria such as\nefficiency and scalability, which are vital in practice. To tackle this\nchallenge, we propose to model the heuristic search as a multi-objective\noptimization problem and consider introducing additional practical criteria\nbeyond optimal performance. Due to the complexity of the search space,\nconventional multi-objective optimization methods struggle to effectively\nhandle LLM-based multi-objective heuristic search. We propose the first\nLLM-based multi-objective heuristic search framework, Multi-objective Evolution\nof Heuristic (MEoH), which integrates LLMs in a zero-shot manner to generate a\nnon-dominated set of heuristics to meet multiple design criteria. We design a\nnew dominance-dissimilarity mechanism for effective population management and\nselection, which incorporates both code dissimilarity in the search space and\ndominance in the objective space. MEoH is demonstrated in two well-known\ncombinatorial optimization problems: the online Bin Packing Problem (BPP) and\nthe Traveling Salesman Problem (TSP). The results indicate that a variety of\nelite heuristics are automatically generated in a single run, offering more\ntrade-off options than the existing methods. It successfully achieves\ncompetitive or superior performance while improving efficiency up to 10 times.\nMoreover, we also observe that the multi-objective search introduces novel\ninsights into heuristic design and leads to the discovery of diverse\nheuristics.\n", "date": "2024-09-25", "categories": []} {"id": "2501.02977", "title": "CAMP: Collaborative Attention Model with Profiles for Vehicle Routing\n Problems", "abstract": " The profiled vehicle routing problem (PVRP) is a generalization of the\nheterogeneous capacitated vehicle routing problem (HCVRP) in which the\nobjective is to optimize the routes of vehicles to serve client demands subject\nto different vehicle profiles, with each having a preference or constraint on a\nper-client basis. While existing learning methods have shown promise for\nsolving the HCVRP in real-time, no learning method exists to solve the more\npractical and challenging PVRP. In this paper, we propose a Collaborative\nAttention Model with Profiles (CAMP), a novel approach that learns efficient\nsolvers for PVRP using multi-agent reinforcement learning. CAMP employs a\nspecialized attention-based encoder architecture to embed profiled client\nembeddings in parallel for each vehicle profile. We design a communication\nlayer between agents for collaborative decision-making across profiled\nembeddings at each decoding step and a batched pointer mechanism to attend to\nthe profiled embeddings to evaluate the likelihood of the next actions. We\nevaluate CAMP on two variants of PVRPs: PVRP with preferences, which explicitly\ninfluence the reward function, and PVRP with zone constraints with different\nnumbers of agents and clients, demonstrating that our learned solvers achieve\ncompetitive results compared to both classical state-of-the-art neural\nmulti-agent models in terms of solution quality and computational efficiency.\nWe make our code openly available at https://github.com/ai4co/camp.\n", "date": "2025-01-06", "categories": []} {"id": "2502.00018", "title": "An Expectation-Maximization Algorithm-based Autoregressive Model for the\n Fuzzy Job Shop Scheduling Problem", "abstract": " The fuzzy job shop scheduling problem (FJSSP) emerges as an innovative\nextension to the job shop scheduling problem (JSSP), incorporating a layer of\nuncertainty that aligns the problem more closely with the complexities of\nreal-world manufacturing environments. This improvement increases the\ncomputational complexity of deriving the solution while improving its\napplicability. In the domain of deterministic scheduling, neural combinatorial\noptimization (NCO) has recently demonstrated remarkable efficacy. However, its\napplication to the realm of fuzzy scheduling has been relatively unexplored.\nThis paper aims to bridge this gap by investigating the feasibility of\nemploying neural networks to assimilate and process fuzzy information for the\nresolution of FJSSP, thereby leveraging the advancements in NCO to enhance\nfuzzy scheduling methodologies. To achieve this, we approach the FJSSP as a\ngenerative task and introduce an expectation-maximization algorithm-based\nautoregressive model (EMARM) to address it. During training, our model\nalternates between generating scheduling schemes from given instances (E-step)\nand adjusting the autoregressive model weights based on these generated schemes\n(M-step). This novel methodology effectively navigates around the substantial\nhurdle of obtaining ground-truth labels, which is a prevalent issue in NCO\nframeworks. In testing, the experimental results demonstrate the superior\ncapability of EMARM in addressing the FJSSP, showcasing its effectiveness and\npotential for practical applications in fuzzy scheduling.\n", "date": "2025-01-11", "categories": []} {"id": "2502.00767", "title": "Learning-Based TSP-Solvers Tend to Be Overly Greedy", "abstract": " Deep learning has shown significant potential in solving combinatorial\noptimization problems such as the Euclidean traveling salesman problem (TSP).\nHowever, most training and test instances for existing TSP algorithms are\ngenerated randomly from specific distributions like uniform distribution. This\nhas led to a lack of analysis and understanding of the performance of deep\nlearning algorithms in out-of-distribution (OOD) generalization scenarios,\nwhich has a close relationship with the worst-case performance in the\ncombinatorial optimization field. For data-driven algorithms, the statistical\nproperties of randomly generated datasets are critical. This study constructs a\nstatistical measure called nearest-neighbor density to verify the asymptotic\nproperties of randomly generated datasets and reveal the greedy behavior of\nlearning-based solvers, i.e., always choosing the nearest neighbor nodes to\nconstruct the solution path. Based on this statistical measure, we develop\ninterpretable data augmentation methods that rely on distribution shifts or\ninstance perturbations and validate that the performance of the learning-based\nsolvers degenerates much on such augmented data. Moreover, fine-tuning\nlearning-based solvers with augmented data further enhances their\ngeneralization abilities. In short, we decipher the limitations of\nlearning-based TSP solvers tending to be overly greedy, which may have profound\nimplications for AI-empowered combinatorial optimization solvers.\n", "date": "2025-02-02", "categories": []} {"id": "2501.08603", "title": "Monte Carlo Tree Search for Comprehensive Exploration in LLM-Based\n Automatic Heuristic Design", "abstract": " Handcrafting heuristics for solving complex optimization tasks (e.g., route\nplanning and task allocation) is a common practice but requires extensive\ndomain knowledge. Recently, Large Language Model (LLM)-based automatic\nheuristic design (AHD) methods have shown promise in generating high-quality\nheuristics without manual interventions. Existing LLM-based AHD methods employ\na population to maintain a fixed number of top-performing LLM-generated\nheuristics and introduce evolutionary computation (EC) to iteratively enhance\nthe population. However, these population-based procedures cannot fully develop\nthe potential of each heuristic and are prone to converge into local optima. To\nmore comprehensively explore the space of heuristics, this paper proposes to\nuse Monte Carlo Tree Search (MCTS) for LLM-based heuristic evolution. The\nproposed MCTS-AHD method organizes all LLM-generated heuristics in a tree\nstructure and can better develop the potential of temporarily underperforming\nheuristics. In experiments, MCTS-AHD delivers significantly higher-quality\nheuristics on various complex tasks. Our code is available.\n", "date": "2025-01-15", "categories": []} {"id": "2501.19063", "title": "Optimizing Job Allocation using Reinforcement Learning with Graph Neural\n Networks", "abstract": " Efficient job allocation in complex scheduling problems poses significant\nchallenges in real-world applications. In this report, we propose a novel\napproach that leverages the power of Reinforcement Learning (RL) and Graph\nNeural Networks (GNNs) to tackle the Job Allocation Problem (JAP). The JAP\ninvolves allocating a maximum set of jobs to available resources while\nconsidering several constraints. Our approach enables learning of adaptive\npolicies through trial-and-error interactions with the environment while\nexploiting the graph-structured data of the problem. By leveraging RL, we\neliminate the need for manual annotation, a major bottleneck in supervised\nlearning approaches. Experimental evaluations on synthetic and real-world data\ndemonstrate the effectiveness and generalizability of our proposed approach,\noutperforming baseline algorithms and showcasing its potential for optimizing\njob allocation in complex scheduling problems.\n", "date": "2025-01-31", "categories": []} {"id": "2501.17991", "title": "Investigating the Monte-Carlo Tree Search Approach for the Job Shop\n Scheduling Problem", "abstract": " The Job Shop Scheduling Problem (JSSP) is a well-known optimization problem\nin manufacturing, where the goal is to determine the optimal sequence of jobs\nacross different machines to minimize a given objective. In this work, we focus\non minimising the weighted sum of job completion times. We explore the\npotential of Monte Carlo Tree Search (MCTS), a heuristic-based reinforcement\nlearning technique, to solve large-scale JSSPs, especially those with\nrecirculation. We propose several Markov Decision Process (MDP) formulations to\nmodel the JSSP for the MCTS algorithm. In addition, we introduce a new\nsynthetic benchmark derived from real manufacturing data, which captures the\ncomplexity of large, non-rectangular instances often encountered in practice.\nOur experimental results show that MCTS effectively produces good-quality\nsolutions for large-scale JSSP instances, outperforming our constraint\nprogramming approach.\n", "date": "2025-01-29", "categories": []} {"id": "2501.17170", "title": "Benchmarking Randomized Optimization Algorithms on Binary, Permutation,\n and Combinatorial Problem Landscapes", "abstract": " In this paper, we evaluate the performance of four randomized optimization\nalgorithms: Randomized Hill Climbing (RHC), Simulated Annealing (SA), Genetic\nAlgorithms (GA), and MIMIC (Mutual Information Maximizing Input Clustering),\nacross three distinct types of problems: binary, permutation, and\ncombinatorial. We systematically compare these algorithms using a set of\nbenchmark fitness functions that highlight the specific challenges and\nrequirements of each problem category. Our study analyzes each algorithm's\neffectiveness based on key performance metrics, including solution quality,\nconvergence speed, computational cost, and robustness. Results show that while\nMIMIC and GA excel in producing high-quality solutions for binary and\ncombinatorial problems, their computational demands vary significantly. RHC and\nSA, while computationally less expensive, demonstrate limited performance in\ncomplex problem landscapes. The findings offer valuable insights into the\ntrade-offs between different optimization strategies and provide practical\nguidance for selecting the appropriate algorithm based on the type of problems,\naccuracy requirements, and computational constraints.\n", "date": "2025-01-21", "categories": []} {"id": "2501.17188", "title": "Letters, Colors, and Words: Constructing the Ideal Building Blocks Set", "abstract": " Define a building blocks set to be a collection of n cubes (each with six\nsides) where each side is assigned one letter and one color from a palette of m\ncolors. We propose a novel problem of assigning letters and colors to each face\nso as to maximize the number of words one can spell from a chosen dataset that\nare either mono words, all letters have the same color, or rainbow words, all\nletters have unique colors. We explore this problem considering a chosen set of\nEnglish words, up to six letters long, from a typical vocabulary of a US\nAmerican 14 year old and explore the problem when n=6 and m=6, with the added\nrestriction that each color appears exactly once on the cube. The problem is\nintractable, as the size of the solution space makes a brute force approach\ncomputationally infeasible. Therefore we aim to solve this problem using random\nsearch, simulated annealing, two distinct tree search approaches (greedy and\nbest-first), and a genetic algorithm. To address this, we explore a range of\noptimization techniques: random search, simulated annealing, two distinct tree\nsearch methods (greedy and best-first), and a genetic algorithm. Additionally,\nwe attempted to implement a reinforcement learning approach; however, the model\nfailed to converge to viable solutions within the problem's constraints. Among\nthese methods, the genetic algorithm delivered the best performance, achieving\na total of 2846 mono and rainbow words.\n", "date": "2025-01-26", "categories": []} {"id": "2501.17725", "title": "Using Code Generation to Solve Open Instances of Combinatorial Design\n Problems", "abstract": " The Handbook of Combinatorial Designs catalogs many types of combinatorial\ndesigns, together with lists of open instances for which existence has not yet\nbeen determined. We develop a constructive protocol CPro1, which uses Large\nLanguage Models (LLMs) to generate code that constructs combinatorial designs\nand resolves some of these open instances. The protocol starts from a\ndefinition of a particular type of design, and a verifier that reliably\nconfirms whether a proposed design is valid. The LLM selects strategies and\nimplements them in code, and scaffolding provides automated hyperparameter\ntuning and execution feedback using the verifier. Most generated code fails,\nbut by generating many candidates, the protocol automates exploration of a\nvariety of standard methods (e.g. simulated annealing, genetic algorithms) and\nexperimentation with variations (e.g. cost functions) to find successful\napproaches. Testing on 16 different types of designs, CPro1 constructs\nsolutions to open instances for 6 of them: Symmetric and Skew Weighing\nMatrices, Equidistant Permutation Arrays, Packing Arrays, Balanced Ternary\nDesigns, and Florentine Rectangles.\n", "date": "2025-01-29", "categories": []} {"id": "2501.16986", "title": "Generative quantum combinatorial optimization by means of a novel\n conditional generative quantum eigensolver", "abstract": " Quantum computing is entering a transformative phase with the emergence of\nlogical quantum processors, which hold the potential to tackle complex problems\nbeyond classical capabilities. While significant progress has been made,\napplying quantum algorithms to real-world problems remains challenging. Hybrid\nquantum-classical techniques have been explored to bridge this gap, but they\noften face limitations in expressiveness, trainability, or scalability. In this\nwork, we introduce conditional Generative Quantum Eigensolver\n(conditional-GQE), a context-aware quantum circuit generator powered by an\nencoder-decoder Transformer. Focusing on combinatorial optimization, we train\nour generator for solving problems with up to 10 qubits, exhibiting nearly\nperfect performance on new problems. By leveraging the high expressiveness and\nflexibility of classical generative models, along with an efficient\npreference-based training scheme, conditional-GQE provides a generalizable and\nscalable framework for quantum circuit generation. Our approach advances hybrid\nquantum-classical computing and contributes to accelerate the transition toward\nfault-tolerant quantum computing.\n", "date": "2025-01-28", "categories": []} {"id": "2405.08604", "title": "Context-aware Diversity Enhancement for Neural Multi-Objective\n Combinatorial Optimization", "abstract": " Multi-objective combinatorial optimization (MOCO) problems are prevalent in\nvarious real-world applications. Most existing neural MOCO methods rely on\nproblem decomposition to transform an MOCO problem into a series of\nsinge-objective combinatorial optimization (SOCO) problems and train attention\nmodels based on a single-step and deterministic greedy rollout. However,\ninappropriate decomposition and undesirable short-sighted behaviors of previous\nmethods tend to induce a decline in diversity. To address the above limitation,\nwe design a Context-aware Diversity Enhancement algorithm named CDE, which\ncasts the neural MOCO problems as conditional sequence modeling via\nautoregression (node-level context awareness) and establishes a direct\nrelationship between the mapping of preferences and diversity indicator of\nreward based on hypervolume expectation maximization (solution-level context\nawareness). Based on the solution-level context awareness, we further propose a\nhypervolume residual update strategy to enable the Pareto attention model to\ncapture both local and non-local information of the Pareto set/front. The\nproposed CDE can effectively and efficiently grasp the context information,\nresulting in diversity enhancement. Experimental results on three classic MOCO\nproblems demonstrate that our CDE outperforms several state-of-the-art\nbaselines.\n", "date": "2024-05-14", "categories": []} {"id": "2501.14775", "title": "Hybrid Firefly-Genetic Algorithm for Single and Multi-dimensional 0-1\n Knapsack Problems", "abstract": " This paper addresses the challenges faced by algorithms, such as the Firefly\nAlgorithm (FA) and the Genetic Algorithm (GA), in constrained optimization\nproblems. While both algorithms perform well for unconstrained problems, their\neffectiveness diminishes when constraints are introduced due to limitations in\nexploration, exploitation, and constraint handling. To overcome these\nchallenges, a hybrid FAGA algorithm is proposed, combining the strengths of\nboth algorithms. The hybrid algorithm is validated by solving unconstrained\nbenchmark functions and constrained optimization problems, including design\nengineering problems and combinatorial problems such as the 0-1 Knapsack\nProblem. The proposed algorithm delivers improved solution accuracy and\ncomputational efficiency compared to conventional optimization algorithm. This\npaper outlines the development and structure of the hybrid algorithm and\ndemonstrates its effectiveness in handling complex optimization problems.\n", "date": "2024-12-31", "categories": []} {"id": "2501.10243", "title": "Random-Key Algorithms for Optimizing Integrated Operating Room\n Scheduling", "abstract": " Efficient surgery room scheduling is essential for hospital efficiency,\npatient satisfaction, and resource utilization. This study addresses this\nchallenge by introducing a novel concept of Random-Key Optimizer (RKO),\nrigorously tested on literature and new, real-world inspired instances. Our\ncombinatorial optimization problem incorporates multi-room scheduling,\nequipment scheduling, and complex availability constraints for rooms, patients,\nand surgeons, facilitating rescheduling and enhancing operational flexibility.\nThe RKO approach represents solutions as points in a continuous space, which\nare then mapped in the problem solution space via a deterministic function\nknown as a decoder. The core idea is to operate metaheuristics and heuristics\nin the random-key space, unaware of the original solution space. We design the\nBiased Random-Key Genetic Algorithm with $Q$-Learning, Simulated Annealing, and\nIterated Local Search for use within an RKO framework, employing a single\ndecoder function. The proposed metaheuristics are complemented by lower-bound\nformulations, providing optimal gaps for evaluating the effectiveness of the\nheuristic results. Our results demonstrate significant lower and upper bounds\nimprovements for the literature instances, notably proving one optimal result.\nFurthermore, the best-proposed metaheuristic efficiently generates schedules\nfor the newly introduced instances, even in highly constrained scenarios. This\nresearch offers valuable insights and practical solutions for improving surgery\nscheduling processes, offering tangible benefits to hospitals by optimising\nresource allocation, reducing patient wait times, and enhancing overall\noperational efficiency.\n", "date": "2025-01-17", "categories": []} {"id": "2501.14394", "title": "Reinforcement Learning for Efficient Returns Management", "abstract": " In retail warehouses, returned products are typically placed in an\nintermediate storage until a decision regarding further shipment to stores is\nmade. The longer products are held in storage, the higher the inefficiency and\ncosts of the returns management process, since enough storage area has to be\nprovided and maintained while the products are not placed for sale. To reduce\nthe average product storage time, we consider an alternative solution where\nreallocation decisions for products can be made instantly upon their arrival in\nthe warehouse allowing only a limited number of products to still be stored\nsimultaneously. We transfer the problem to an online multiple knapsack problem\nand propose a novel reinforcement learning approach to pack the items\n(products) into the knapsacks (stores) such that the overall value (expected\nrevenue) is maximized. Empirical evaluations on simulated data demonstrate\nthat, compared to the usual offline decision procedure, our approach comes with\na performance gap of only 3% while significantly reducing the average storage\ntime of a product by 96%.\n", "date": "2025-01-24", "categories": []} {"id": "2501.13767", "title": "An Efficient Diffusion-based Non-Autoregressive Solver for Traveling\n Salesman Problem", "abstract": " Recent advances in neural models have shown considerable promise in solving\nTraveling Salesman Problems (TSPs) without relying on much hand-crafted\nengineering. However, while non-autoregressive (NAR) approaches benefit from\nfaster inference through parallelism, they typically deliver solutions of\ninferior quality compared to autoregressive ones. To enhance the solution\nquality while maintaining fast inference, we propose DEITSP, a diffusion model\nwith efficient iterations tailored for TSP that operates in a NAR manner.\nFirstly, we introduce a one-step diffusion model that integrates the controlled\ndiscrete noise addition process with self-consistency enhancement, enabling\noptimal solution prediction through simultaneous denoising of multiple\nsolutions. Secondly, we design a dual-modality graph transformer to bolster the\nextraction and fusion of features from node and edge modalities, while further\naccelerating the inference with fewer layers. Thirdly, we develop an efficient\niterative strategy that alternates between adding and removing noise to improve\nexploration compared to previous diffusion methods. Additionally, we devise a\nscheduling framework to progressively refine the solution space by adjusting\nnoise levels, facilitating a smooth search for optimal solutions. Extensive\nexperiments on real-world and large-scale TSP instances demonstrate that DEITSP\nperforms favorably against existing neural approaches in terms of solution\nquality, inference latency, and generalization ability. Our code is available\nat $\\href{https://github.com/DEITSP/DEITSP}{https://github.com/DEITSP/DEITSP}$.\n", "date": "2025-01-23", "categories": []} {"id": "2407.00312", "title": "UDC: A Unified Neural Divide-and-Conquer Framework for Large-Scale\n Combinatorial Optimization Problems", "abstract": " Single-stage neural combinatorial optimization solvers have achieved\nnear-optimal results on various small-scale combinatorial optimization (CO)\nproblems without requiring expert knowledge. However, these solvers exhibit\nsignificant performance degradation when applied to large-scale CO problems.\nRecently, two-stage neural methods motivated by divide-and-conquer strategies\nhave shown efficiency in addressing large-scale CO problems. Nevertheless, the\nperformance of these methods highly relies on problem-specific heuristics in\neither the dividing or the conquering procedure, which limits their\napplicability to general CO problems. Moreover, these methods employ separate\ntraining schemes and ignore the interdependencies between the dividing and\nconquering strategies, often leading to sub-optimal solutions. To tackle these\ndrawbacks, this article develops a unified neural divide-and-conquer framework\n(i.e., UDC) for solving general large-scale CO problems. UDC offers a\nDivide-Conquer-Reunion (DCR) training method to eliminate the negative impact\nof a sub-optimal dividing policy. Employing a high-efficiency Graph Neural\nNetwork (GNN) for global instance dividing and a fixed-length sub-path solver\nfor conquering divided sub-problems, the proposed UDC framework demonstrates\nextensive applicability, achieving superior performance in 10 representative\nlarge-scale CO problems. The code is available at\nhttps://github.com/CIAM-Group/NCO_code/tree/main/single_objective/UDC-Large-scale-CO-master.\n", "date": "2024-06-29", "categories": []} {"id": "2408.07510", "title": "Dominating Set Reconfiguration with Answer Set Programming", "abstract": " The dominating set reconfiguration problem is defined as determining, for a\ngiven dominating set problem and two among its feasible solutions, whether one\nis reachable from the other via a sequence of feasible solutions subject to a\ncertain adjacency relation. This problem is PSPACE-complete in general. The\nconcept of the dominating set is known to be quite useful for analyzing\nwireless networks, social networks, and sensor networks. We develop an approach\nto solve the dominating set reconfiguration problem based on Answer Set\nProgramming (ASP). Our declarative approach relies on a high-level ASP\nencoding, and both the grounding and solving tasks are delegated to an\nASP-based combinatorial reconfiguration solver. To evaluate the effectiveness\nof our approach, we conduct experiments on a newly created benchmark set.\n", "date": "2024-08-14", "categories": []} {"id": "2501.11240", "title": "Fast instance-specific algorithm configuration with graph neural network", "abstract": " Combinatorial optimization (CO) problems are pivotal across various\nindustrial applications, where the speed of solving these problems is crucial.\nImproving the performance of CO solvers across diverse input instances requires\nfine-tuning solver parameters for each instance. However, this tuning process\nis time-consuming, and the time required increases with the number of\ninstances. To address this, a method called instance-specific algorithm\nconfiguration (ISAC) has been devised. This approach involves two main steps:\ntraining and execution. During the training step, features are extracted from\nvarious instances and then grouped into clusters. For each cluster, parameters\nare fine-tuned. This cluster-specific tuning process results in a set of\ngeneralized parameters for instances belonging to each class. In the execution\nstep, features are extracted from an unknown instance to determine its cluster,\nand the corresponding pre-tuned parameters are applied. Generally, the running\ntime of a solver is evaluated by the time to solution ($TTS$). However, methods\nlike ISAC require preprocessing. Therefore, the total execution time is\n$T_{tot}=TTS+T_{tune}$, where $T_{tune}$ represents the tuning time. While the\ngoal is to minimize $T_{tot}$, it is important to note that extracting features\nin the ISAC method requires a certain amount of computational time. The\nextracting features include summary statistics of the solver execution logs,\nwhich takes several 10 seconds. This research presents a method to\nsignificantly reduce the time of the ISAC execution step by streamlining\nfeature extraction and class determination with a graph neural network.\nExperimental results show that $T_{tune}$ in the execution step, which take\nseveral 10 seconds in the original ISAC manner, could be reduced to\nsub-seconds.\n", "date": "2025-01-20", "categories": []} {"id": "2501.11968", "title": "Bridging Visualization and Optimization: Multimodal Large Language\n Models on Graph-Structured Combinatorial Optimization", "abstract": " Graph-structured combinatorial challenges are inherently difficult due to\ntheir nonlinear and intricate nature, often rendering traditional computational\nmethods ineffective or expensive. However, these challenges can be more\nnaturally tackled by humans through visual representations that harness our\ninnate ability for spatial reasoning. In this study, we propose transforming\ngraphs into images to preserve their higher-order structural features\naccurately, revolutionizing the representation used in solving graph-structured\ncombinatorial tasks. This approach allows machines to emulate human-like\nprocessing in addressing complex combinatorial challenges. By combining the\ninnovative paradigm powered by multimodal large language models (MLLMs) with\nsimple search techniques, we aim to develop a novel and effective framework for\ntackling such problems. Our investigation into MLLMs spanned a variety of\ngraph-based tasks, from combinatorial problems like influence maximization to\nsequential decision-making in network dismantling, as well as addressing six\nfundamental graph-related issues. Our findings demonstrate that MLLMs exhibit\nexceptional spatial intelligence and a distinctive capability for handling\nthese problems, significantly advancing the potential for machines to\ncomprehend and analyze graph-structured data with a depth and intuition akin to\nhuman cognition. These results also imply that integrating MLLMs with simple\noptimization strategies could form a novel and efficient approach for\nnavigating graph-structured combinatorial challenges without complex\nderivations, computationally demanding training and fine-tuning.\n", "date": "2025-01-21", "categories": []} {"id": "2501.08565", "title": "DualOpt: A Dual Divide-and-Optimize Algorithm for the Large-scale\n Traveling Salesman Problem", "abstract": " This paper proposes a dual divide-and-optimize algorithm (DualOpt) for\nsolving the large-scale traveling salesman problem (TSP). DualOpt combines two\ncomplementary strategies to improve both solution quality and computational\nefficiency. The first strategy is a grid-based divide-and-conquer procedure\nthat partitions the TSP into smaller sub-problems, solving them in parallel and\niteratively refining the solution by merging nodes and partial routes. The\nprocess continues until only one grid remains, yielding a high-quality initial\nsolution. The second strategy involves a path-based divide-and-optimize\nprocedure that further optimizes the solution by dividing it into sub-paths,\noptimizing each using a neural solver, and merging them back to progressively\nimprove the overall solution. Extensive experiments conducted on two groups of\nTSP benchmark instances, including randomly generated instances with up to\n100,000 nodes and real-world datasets from TSPLIB, demonstrate the\neffectiveness of DualOpt. The proposed DualOpt achieves highly competitive\nresults compared to 10 state-of-the-art algorithms in the literature. In\nparticular, DualOpt achieves an improvement gap up to 1.40% for the largest\ninstance TSP100K with a remarkable 104x speed-up over the leading heuristic\nsolver LKH3. Additionally, DualOpt demonstrates strong generalization on TSPLIB\nbenchmarks, confirming its capability to tackle diverse real-world TSP\napplications.\n", "date": "2025-01-15", "categories": []} {"id": "2403.04917", "title": "A Mixed-Integer Conic Program for the Moving-Target Traveling Salesman\n Problem based on a Graph of Convex Sets", "abstract": " This paper introduces a new formulation that finds the optimum for the\nMoving-Target Traveling Salesman Problem (MT-TSP), which seeks to find a\nshortest path for an agent, that starts at a depot, visits a set of moving\ntargets exactly once within their assigned time-windows, and returns to the\ndepot. The formulation relies on the key idea that when the targets move along\nlines, their trajectories become convex sets within the space-time coordinate\nsystem. The problem then reduces to finding the shortest path within a graph of\nconvex sets, subject to some speed constraints. We compare our formulation with\nthe current state-of-the-art Mixed Integer Conic Program (MICP) solver for the\nMT-TSP. The experimental results show that our formulation outperforms the MICP\nfor instances with up to 20 targets, with up to two orders of magnitude\nreduction in runtime, and up to a 60\\% tighter optimality gap. We also show\nthat the solution cost from the convex relaxation of our formulation provides\nsignificantly tighter lower bounds for the MT-TSP than the ones from the MICP.\n", "date": "2024-03-07", "categories": []} {"id": "2501.07896", "title": "Anytime Cooperative Implicit Hitting Set Solving", "abstract": " The Implicit Hitting Set (HS) approach has shown to be very effective for\nMaxSAT, Pseudo-boolean optimization and other boolean frameworks. Very\nrecently, it has also shown its potential in the very similar Weighted CSP\nframework by means of the so-called cost-function merging. The original\nformulation of the HS approach focuses on obtaining increasingly better lower\nbounds (HS-lb). However, and as shown for Pseudo-Boolean Optimization, this\napproach can also be adapted to compute increasingly better upper bounds\n(HS-ub). In this paper we consider both HS approaches and show how they can be\neasily combined in a multithread architecture where cores discovered by either\ncomponent are available by the other which, interestingly, generates synergy\nbetween them. We show that the resulting algorithm (HS-lub) is consistently\nsuperior to either HS-lb and HS-ub in isolation. Most importantly, HS-lub has\nan effective anytime behaviour with which the optimality gap is reduced during\nthe execution. We tested our approach on the Weighted CSP framework and show on\nthree different benchmarks that our very simple implementation sometimes\noutperforms the parallel hybrid best-first search implementation of the far\nmore developed state-of-the-art Toulbar2.\n", "date": "2025-01-14", "categories": []} {"id": "2501.08020", "title": "Cooperative Patrol Routing: Optimizing Urban Crime Surveillance through\n Multi-Agent Reinforcement Learning", "abstract": " The effective design of patrol strategies is a difficult and complex problem,\nespecially in medium and large areas. The objective is to plan, in a\ncoordinated manner, the optimal routes for a set of patrols in a given area, in\norder to achieve maximum coverage of the area, while also trying to minimize\nthe number of patrols. In this paper, we propose a multi-agent reinforcement\nlearning (MARL) model, based on a decentralized partially observable Markov\ndecision process, to plan unpredictable patrol routes within an urban\nenvironment represented as an undirected graph. The model attempts to maximize\na target function that characterizes the environment within a given time frame.\nOur model has been tested to optimize police patrol routes in three\nmedium-sized districts of the city of Malaga. The aim was to maximize\nsurveillance coverage of the most crime-prone areas, based on actual crime data\nin the city. To address this problem, several MARL algorithms have been\nstudied, and among these the Value Decomposition Proximal Policy Optimization\n(VDPPO) algorithm exhibited the best performance. We also introduce a novel\nmetric, the coverage index, for the evaluation of the coverage performance of\nthe routes generated by our model. This metric is inspired by the predictive\naccuracy index (PAI), which is commonly used in criminology to detect hotspots.\nUsing this metric, we have evaluated the model under various scenarios in which\nthe number of agents (or patrols), their starting positions, and the level of\ninformation they can observe in the environment have been modified. Results\nshow that the coordinated routes generated by our model achieve a coverage of\nmore than $90\\%$ of the $3\\%$ of graph nodes with the highest crime incidence,\nand $65\\%$ for $20\\%$ of these nodes; $3\\%$ and $20\\%$ represent the coverage\nstandards for police resource allocation.\n", "date": "2025-01-14", "categories": []} {"id": "2207.11132", "title": "Proactive Distributed Emergency Response with Heterogeneous Tasks\n Allocation", "abstract": " Traditionally, traffic incident management (TIM) programs coordinate the\ndeployment of emergency resources to immediate incident requests without\naccommodating the interdependencies on incident evolutions in the environment.\nHowever, ignoring inherent interdependencies on the evolution of incidents in\nthe environment while making current deployment decisions is shortsighted, and\nthe resulting naive deployment strategy can significantly worsen the overall\nincident delay impact on the network. The interdependencies on incident\nevolution in the environment, including those between incident occurrences, and\nthose between resource availability in near-future requests and the anticipated\nduration of the immediate incident request, should be considered through a\nlook-ahead model when making current-stage deployment decisions. This study\ndevelops a new proactive framework based on the distributed constraint\noptimization problem (DCOP) to address the above limitations, overcoming\nconventional TIM models that cannot accommodate the dependencies in the TIM\nproblem. Furthermore, the optimization objective is formulated to incorporate\nUnmanned Aerial Vehicles (UAVs). The UAVs' role in TIM includes exploring\nuncertain traffic conditions, detecting unexpected events, and augmenting\ninformation from roadway traffic sensors. Robustness analysis of our model for\nmultiple TIM scenarios shows satisfactory performance using local search\nexploration heuristics. Overall, our model reports a significant reduction in\ntotal incident delay compared to conventional TIM models. With UAV support, we\ndemonstrate a further decrease in the total incident delay ranging between 5%\nand 45% for the different number of incidents. UAV's active sensing can shorten\nresponse time of emergency vehicles, and a reduction in uncertainties\nassociated with the estimated incident delay impact.\n", "date": "2022-07-16", "categories": []} {"id": "2412.15908", "title": "Speedup Techniques for Switchable Temporal Plan Graph Optimization", "abstract": " Multi-Agent Path Finding (MAPF) focuses on planning collision-free paths for\nmultiple agents. However, during the execution of a MAPF plan, agents may\nencounter unexpected delays, which can lead to inefficiencies, deadlocks, or\neven collisions. To address these issues, the Switchable Temporal Plan Graph\nprovides a framework for finding an acyclic Temporal Plan Graph with the\nminimum execution cost under delays, ensuring deadlock- and collision-free\nexecution. Unfortunately, existing optimal algorithms, such as Mixed Integer\nLinear Programming and Graph-Based Switchable Edge Search (GSES), are often too\nslow for practical use. This paper introduces Improved GSES, which\nsignificantly accelerates GSES through four speedup techniques: stronger\nadmissible heuristics, edge grouping, prioritized branching, and incremental\nimplementation. Experiments conducted on four different map types with varying\nnumbers of agents demonstrate that Improved GSES consistently achieves over\ntwice the success rate of GSES and delivers up to a 30-fold speedup on\ninstances where both methods successfully find solutions.\n", "date": "2024-12-20", "categories": []} {"id": "2501.06506", "title": "Resource Allocation under the Latin Square Constraint", "abstract": " A Latin square is an $n \\times n$ matrix filled with $n$ distinct symbols,\neach of which appears exactly once in each row and exactly once in each column.\nWe introduce a problem of allocating $n$ indivisible items among $n$ agents\nover $n$ rounds while satisfying the Latin square constraint. This constraint\nensures that each agent receives no more than one item per round and receives\neach item at most once. Each agent has an additive valuation on the item--round\npairs. Real-world applications like scheduling, resource management, and\nexperimental design require the Latin square constraint to satisfy fairness or\nbalancedness in allocation. Our goal is to find a partial or complete\nallocation that maximizes the sum of the agents' valuations (utilitarian social\nwelfare) or the minimum of the agents' valuations (egalitarian social welfare).\nFor the problem of maximizing utilitarian social welfare, we prove NP-hardness\neven when the valuations are binary additive. We then provide $(1-1/e)$ and\n$(1-1/e)/4$-approximation algorithms for partial and complete settings,\nrespectively. Additionally, we present fixed-parameter tractable (FPT)\nalgorithms with respect to the order of Latin square and the optimum value for\nboth partial and complete settings. For the problem of maximizing egalitarian\nsocial welfare, we establish that deciding whether the optimum value is at most\n$1$ or at least $2$ is NP-hard for both the partial and complete settings, even\nwhen the valuations are binary. Furthermore, we demonstrate that checking the\nexistence of a complete allocation that satisfies each of envy-free,\nproportional, equitable, envy-free up to any good, proportional up to any good,\nor equitable up to any good is NP-hard, even when the valuations are identical.\n", "date": "2025-01-11", "categories": []} {"id": "2501.07432", "title": "Empirical Evaluation of the Implicit Hitting Set Approach for Weighted\n CSPs", "abstract": " SAT technology has proven to be surprisingly effective in a large variety of\ndomains. However, for the Weighted CSP problem dedicated algorithms have always\nbeen superior. One approach not well-studied so far is the use of SAT in\nconjunction with the Implicit Hitting Set approach. In this work, we explore\nsome alternatives to the existing algorithm of reference. The alternatives,\nmostly borrowed from related boolean frameworks, consider trade-offs for the\ntwo main components of the IHS approach: the computation of low-cost hitting\nvectors, and their transformation into high-cost cores. For each one, we\npropose 4 levels of intensity. Since we also test the usefulness of cost\nfunction merging, our experiments consider 32 different implementations. Our\nempirical study shows that for WCSP it is not easy to identify the best\nalternative. Nevertheless, the cost-function merging encoding and extracting\nmaximal cores seems to be a robust approach.\n", "date": "2025-01-13", "categories": []} {"id": "2501.05673", "title": "Network Diffuser for Placing-Scheduling Service Function Chains with\n Inverse Demonstration", "abstract": " Network services are increasingly managed by considering chained-up virtual\nnetwork functions and relevant traffic flows, known as the Service Function\nChains (SFCs). To deal with sequential arrivals of SFCs in an online fashion,\nwe must consider two closely-coupled problems - an SFC placement problem that\nmaps SFCs to servers/links in the network and an SFC scheduling problem that\ndetermines when each SFC is executed. Solving the whole SFC problem targeting\nthese two optimizations jointly is extremely challenging. In this paper, we\npropose a novel network diffuser using conditional generative modeling for this\nSFC placing-scheduling optimization. Recent advances in generative AI and\ndiffusion models have made it possible to generate high-quality images/videos\nand decision trajectories from language description. We formulate the SFC\noptimization as a problem of generating a state sequence for planning and\nperform graph diffusion on the state trajectories to enable extraction of SFC\ndecisions, with SFC optimization constraints and objectives as conditions. To\naddress the lack of demonstration data due to NP-hardness and exponential\nproblem space of the SFC optimization, we also propose a novel and somewhat\nmaverick approach -- Rather than solving instances of this difficult\noptimization, we start with randomly-generated solutions as input, and then\ndetermine appropriate SFC optimization problems that render these solutions\nfeasible. This inverse demonstration enables us to obtain sufficient expert\ndemonstrations, i.e., problem-solution pairs, through further optimization. In\nour numerical evaluations, the proposed network diffuser outperforms learning\nand heuristic baselines, by $\\sim$20\\% improvement in SFC reward and $\\sim$50\\%\nreduction in SFC waiting time and blocking rate.\n", "date": "2025-01-10", "categories": []} {"id": "2501.05845", "title": "Annealing Machine-assisted Learning of Graph Neural Network for\n Combinatorial Optimization", "abstract": " While Annealing Machines (AM) have shown increasing capabilities in solving\ncomplex combinatorial problems, positioning themselves as a more immediate\nalternative to the expected advances of future fully quantum solutions, there\nare still scaling limitations. In parallel, Graph Neural Networks (GNN) have\nbeen recently adapted to solve combinatorial problems, showing competitive\nresults and potentially high scalability due to their distributed nature. We\npropose a merging approach that aims at retaining both the accuracy exhibited\nby AMs and the representational flexibility and scalability of GNNs. Our model\nconsiders a compression step, followed by a supervised interaction where\npartial solutions obtained from the AM are used to guide local GNNs from where\nnode feature representations are obtained and combined to initialize an\nadditional GNN-based solver that handles the original graph's target problem.\nIntuitively, the AM can solve the combinatorial problem indirectly by infusing\nits knowledge into the GNN. Experiments on canonical optimization problems show\nthat the idea is feasible, effectively allowing the AM to solve size problems\nbeyond its original limits.\n", "date": "2025-01-10", "categories": []} {"id": "2501.04072", "title": "Multi-armed Bandit and Backbone boost Lin-Kernighan-Helsgaun Algorithm\n for the Traveling Salesman Problems", "abstract": " The Lin-Kernighan-Helsguan (LKH) heuristic is a classic local search\nalgorithm for the Traveling Salesman Problem (TSP). LKH introduces an\n$\\alpha$-value to replace the traditional distance metric for evaluating the\nedge quality, which leads to a significant improvement. However, we observe\nthat the $\\alpha$-value does not make full use of the historical information\nduring the search, and single guiding information often makes LKH hard to\nescape from some local optima. To address the above issues, we propose a novel\nway to extract backbone information during the TSP local search process, which\nis dynamic and can be updated once a local optimal solution is found. We\nfurther propose to combine backbone information, $\\alpha$-value, and distance\nto evaluate the edge quality so as to guide the search. Moreover, we abstract\ntheir different combinations to arms in a multi-armed bandit (MAB) and use an\nMAB model to help the algorithm select an appropriate evaluation metric\ndynamically. Both the backbone information and MAB can provide diverse guiding\ninformation and learn from the search history to suggest the best metric. We\napply our methods to LKH and LKH-3, which is an extension version of LKH that\ncan be used to solve about 40 variant problems of TSP and Vehicle Routing\nProblem (VRP). Extensive experiments show the excellent performance and\ngeneralization capability of our proposed method, significantly improving LKH\nfor TSP and LKH-3 for two representative TSP and VRP variants, the Colored TSP\n(CTSP) and Capacitated VRP with Time Windows (CVRPTW).\n", "date": "2025-01-07", "categories": []} {"id": "2501.04253", "title": "Integrated Offline and Online Learning to Solve a Large Class of\n Scheduling Problems", "abstract": " In this paper, we develop a unified machine learning (ML) approach to predict\nhigh-quality solutions for single-machine scheduling problems with a\nnon-decreasing min-sum objective function with or without release times. Our ML\napproach is novel in three major aspects. First, our approach is developed for\nthe entire class of the aforementioned problems. To achieve this, we exploit\nthe fact that the entire class of the problems considered can be formulated as\na time-indexed formulation in a unified manner. We develop a deep neural\nnetwork (DNN) which uses the cost parameters in the time-indexed formulation as\nthe inputs to effectively predict a continuous solution to this formulation,\nbased on which a feasible discrete solution is easily constructed. The second\nnovel aspect of our approach lies in how the DNN model is trained. In view of\nthe NP-hard nature of the problems, labels (i.e., optimal solutions) are hard\nto generate for training. To overcome this difficulty, we generate and utilize\na set of special instances, for which optimal solutions can be found with\nlittle computational effort, to train the ML model offline. The third novel\nidea we employ in our approach is that we develop an online single-instance\nlearning approach to fine tune the parameters in the DNN for a given online\ninstance, with the goal of generating an improved solution for the given\ninstance. To this end, we develop a feasibility surrogate that approximates the\nobjective value of a given instance as a continuous function of the outputs of\nthe DNN, which then enables us to derive gradients and update the learnable\nparameters in the DNN. Numerical results show that our approach can efficiently\ngenerate high-quality solutions for a variety of single-machine scheduling\nmin-sum problems with up to 1000 jobs.\n", "date": "2025-01-08", "categories": []} {"id": "2405.15119", "title": "Bayesian Optimization of Functions over Node Subsets in Graphs", "abstract": " We address the problem of optimizing over functions defined on node subsets\nin a graph. The optimization of such functions is often a non-trivial task\ngiven their combinatorial, black-box and expensive-to-evaluate nature. Although\nvarious algorithms have been introduced in the literature, most are either\ntask-specific or computationally inefficient and only utilize information about\nthe graph structure without considering the characteristics of the function. To\naddress these limitations, we utilize Bayesian Optimization (BO), a\nsample-efficient black-box solver, and propose a novel framework for\ncombinatorial optimization on graphs. More specifically, we map each $k$-node\nsubset in the original graph to a node in a new combinatorial graph and adopt a\nlocal modeling approach to efficiently traverse the latter graph by\nprogressively sampling its subgraphs using a recursive algorithm. Extensive\nexperiments under both synthetic and real-world setups demonstrate the\neffectiveness of the proposed BO framework on various types of graphs and\noptimization tasks, where its behavior is analyzed in detail with ablation\nstudies.\n", "date": "2024-05-24", "categories": []} {"id": "2501.02803", "title": "Enhancing Lifelong Multi-Agent Path Finding with Cache Mechanism", "abstract": " Multi-Agent Path Finding (MAPF), which focuses on finding collision-free\npaths for multiple robots, is crucial in autonomous warehouse operations.\nLifelong MAPF (L-MAPF), where agents are continuously reassigned new targets\nupon completing their current tasks, offers a more realistic approximation of\nreal-world warehouse scenarios. While cache storage systems can enhance\nefficiency and reduce operational costs, existing approaches primarily rely on\nexpectations and mathematical models, often without adequately addressing the\nchallenges of multi-robot planning and execution. In this paper, we introduce a\nnovel mechanism called Lifelong MAPF with Cache Mechanism (L-MAPF-CM), which\nintegrates high-level cache storage with low-level path planning. We have\ninvolved a new type of map grid called cache for temporary item storage.\nAdditionally, we involved a task assigner (TA) with a locking mechanism to\nbridge the gap between the new cache grid and L-MAPF algorithm. The TA\ndynamically allocates target locations to agents based on their status in\nvarious scenarios. We evaluated L-MAPF-CM using different cache replacement\npolicies and task distributions. L-MAPF-CM has demonstrated performance\nimprovements particularly with high cache hit rates and smooth traffic\nconditions.\n", "date": "2025-01-06", "categories": []} {"id": "2501.03018", "title": "Probably Correct Optimal Stable Matching for Two-Sided Markets Under\n Uncertainty", "abstract": " We consider a learning problem for the stable marriage model under unknown\npreferences for the left side of the market. We focus on the centralized case,\nwhere at each time step, an online platform matches the agents, and obtains a\nnoisy evaluation reflecting their preferences. Our aim is to quickly identify\nthe stable matching that is left-side optimal, rendering this a pure\nexploration problem with bandit feedback. We specifically aim to find Probably\nCorrect Optimal Stable Matchings and present several bandit algorithms to do\nso. Our findings provide a foundational understanding of how to efficiently\ngather and utilize preference information to identify the optimal stable\nmatching in two-sided markets under uncertainty. An experimental analysis on\nsynthetic data complements theoretical results on sample complexities for the\nproposed methods.\n", "date": "2025-01-06", "categories": []} {"id": "2307.09295", "title": "Learning to Select and Rank from Choice-Based Feedback: A Simple Nested\n Approach", "abstract": " We study a ranking and selection problem of learning from choice-based\nfeedback with dynamic assortments. In this problem, a company sequentially\ndisplays a set of items to a population of customers and collects their choices\nas feedback. The only information available about the underlying choice model\nis that the choice probabilities are consistent with some unknown true strict\nranking over the items. The objective is to identify, with the fewest samples,\nthe most preferred item or the full ranking over the items at a high confidence\nlevel. We present novel and simple algorithms for both learning goals. In the\nfirst subproblem regarding best-item identification, we introduce an\nelimination-based algorithm, Nested Elimination (NE). In the more complex\nsubproblem regarding full-ranking identification, we generalize NE and propose\na divide-and-conquer algorithm, Nested Partition (NP). We provide strong\ncharacterizations of both algorithms through instance-specific and\nnon-asymptotic bounds on the sample complexity. This is accomplished using an\nanalytical framework that characterizes the system dynamics through analyzing a\nsequence of multi-dimensional random walks. We also establish a connection\nbetween our nested approach and the information-theoretic lower bounds. We thus\nshow that NE is worst-case asymptotically optimal, and NP is optimal up to a\nconstant factor. Finally, numerical experiments from both synthetic and real\ndata corroborate our theoretical findings.\n", "date": "2023-07-13", "categories": []} {"id": "2501.00307", "title": "Fast and Interpretable Mixed-Integer Linear Program Solving by Learning\n Model Reduction", "abstract": " By exploiting the correlation between the structure and the solution of\nMixed-Integer Linear Programming (MILP), Machine Learning (ML) has become a\npromising method for solving large-scale MILP problems. Existing ML-based MILP\nsolvers mainly focus on end-to-end solution learning, which suffers from the\nscalability issue due to the high dimensionality of the solution space. Instead\nof directly learning the optimal solution, this paper aims to learn a reduced\nand equivalent model of the original MILP as an intermediate step. The reduced\nmodel often corresponds to interpretable operations and is much simpler,\nenabling us to solve large-scale MILP problems much faster than existing\ncommercial solvers. However, current approaches rely only on the optimal\nreduced model, overlooking the significant preference information of all\nreduced models. To address this issue, this paper proposes a preference-based\nmodel reduction learning method, which considers the relative performance\n(i.e., objective cost and constraint feasibility) of all reduced models on each\nMILP instance as preferences. We also introduce an attention mechanism to\ncapture and represent preference information, which helps improve the\nperformance of model reduction learning tasks. Moreover, we propose a SetCover\nbased pruning method to control the number of reduced models (i.e., labels),\nthereby simplifying the learning process. Evaluation on real-world MILP\nproblems shows that 1) compared to the state-of-the-art model reduction ML\nmethods, our method obtains nearly 20% improvement on solution accuracy, and 2)\ncompared to the commercial solver Gurobi, two to four orders of magnitude\nspeedups are achieved.\n", "date": "2024-12-31", "categories": []} {"id": "2501.00852", "title": "Hybridising Reinforcement Learning and Heuristics for Hierarchical\n Directed Arc Routing Problems", "abstract": " The Hierarchical Directed Capacitated Arc Routing Problem (HDCARP) is an\nextension of the Capacitated Arc Routing Problem (CARP), where the arcs of a\ngraph are divided into classes based on their priority. The traversal of these\nclasses is determined by either precedence constraints or a hierarchical\nobjective, resulting in two distinct HDCARP variants. To the best of our\nknowledge, only one matheuristic has been proposed for these variants, but it\nperforms relatively slowly, particularly for large-scale instances (Ha et al.,\n2024). In this paper, we propose a fast heuristic to efficiently address the\ncomputational challenges of HDCARP. Furthermore, we incorporate Reinforcement\nLearning (RL) into our heuristic to effectively guide the selection of local\nsearch operators, resulting in a hybrid algorithm. We name this hybrid\nalgorithm as the Hybrid Reinforcement Learning and Heuristic Algorithm for\nDirected Arc Routing (HRDA). The hybrid algorithm adapts to changes in the\nproblem dynamically, using real-time feedback to improve routing strategies and\nsolution's quality by integrating heuristic methods. Extensive computational\nexperiments on artificial instances demonstrate that this hybrid approach\nsignificantly improves the speed of the heuristic without deteriorating the\nsolution quality. Our source code is publicly available at:\nhttps://github.com/HySonLab/ArcRoute\n", "date": "2025-01-01", "categories": []} {"id": "2501.00884", "title": "Diversity Optimization for Travelling Salesman Problem via Deep\n Reinforcement Learning", "abstract": " Existing neural methods for the Travelling Salesman Problem (TSP) mostly aim\nat finding a single optimal solution. To discover diverse yet high-quality\nsolutions for Multi-Solution TSP (MSTSP), we propose a novel deep reinforcement\nlearning based neural solver, which is primarily featured by an encoder-decoder\nstructured policy. Concretely, on the one hand, a Relativization Filter (RF) is\ndesigned to enhance the robustness of the encoder to affine transformations of\nthe instances, so as to potentially improve the quality of the found solutions.\nOn the other hand, a Multi-Attentive Adaptive Active Search (MA3S) is tailored\nto allow the decoders to strike a balance between the optimality and diversity.\nExperimental evaluations on benchmark instances demonstrate the superiority of\nour method over recent neural baselines across different metrics, and its\ncompetitive performance against state-of-the-art traditional heuristics with\nsignificantly reduced computational time, ranging from $1.3\\times$ to\n$15\\times$ faster. Furthermore, we demonstrate that our method can also be\napplied to the Capacitated Vehicle Routing Problem (CVRP).\n", "date": "2025-01-01", "categories": []} {"id": "2211.17078", "title": "Reinforcement Learning for Multi-Truck Vehicle Routing Problems", "abstract": " Deep reinforcement learning (RL) has been shown to be effective in producing\napproximate solutions to some vehicle routing problems (VRPs), especially when\nusing policies generated by encoder-decoder attention mechanisms. While these\ntechniques have been quite successful for relatively simple problem instances,\nthere are still under-researched and highly complex VRP variants for which no\neffective RL method has been demonstrated. In this work we focus on one such\nVRP variant, which contains multiple trucks and multi-leg routing requirements.\nIn these problems, demand is required to move along sequences of nodes, instead\nof just from a start node to an end node. With the goal of making deep RL a\nviable strategy for real-world industrial-scale supply chain logistics, we\ndevelop new extensions to existing encoder-decoder attention models which allow\nthem to handle multiple trucks and multi-leg routing requirements. Our models\nhave the advantage that they can be trained for a small number of trucks and\nnodes, and then embedded into a large supply chain to yield solutions for\nlarger numbers of trucks and nodes. We test our approach on a real supply chain\nenvironment arising in the operations of Japanese automotive parts manufacturer\nAisin Corporation, and find that our algorithm outperforms Aisin's previous\nbest solution.\n", "date": "2022-11-30", "categories": []} {"id": "2409.18444", "title": "Cost-Aware Dynamic Cloud Workflow Scheduling using Self-Attention and\n Evolutionary Reinforcement Learning", "abstract": " The Cost-aware Dynamic Multi-Workflow Scheduling (CDMWS) in the cloud is a\nkind of cloud workflow management problem, which aims to assign virtual machine\n(VM) instances to execute tasks in workflows so as to minimize the total costs,\nincluding both the penalties for violating Service Level Agreement (SLA) and\nthe VM rental fees. Powered by deep neural networks, Reinforcement Learning\n(RL) methods can construct effective scheduling policies for solving CDMWS\nproblems. Traditional policy networks in RL often use basic feedforward\narchitectures to separately determine the suitability of assigning any VM\ninstances, without considering all VMs simultaneously to learn their global\ninformation. This paper proposes a novel self-attention policy network for\ncloud workflow scheduling (SPN-CWS) that captures global information from all\nVMs. We also develop an Evolution Strategy-based RL (ERL) system to train\nSPN-CWS reliably and effectively. The trained SPN-CWS can effectively process\nall candidate VM instances simultaneously to identify the most suitable VM\ninstance to execute every workflow task. Comprehensive experiments show that\nour method can noticeably outperform several state-of-the-art algorithms on\nmultiple benchmark CDMWS problems.\n", "date": "2024-09-27", "categories": []} {"id": "2412.19066", "title": "FFCG: Effective and Fast Family Column Generation for Solving\n Large-Scale Linear Program", "abstract": " Column Generation (CG) is an effective and iterative algorithm to solve\nlarge-scale linear programs (LP). During each CG iteration, new columns are\nadded to improve the solution of the LP. Typically, CG greedily selects one\ncolumn with the most negative reduced cost, which can be improved by adding\nmore columns at once. However, selecting all columns with negative reduced\ncosts would lead to the addition of redundant columns that do not improve the\nobjective value. Therefore, selecting the appropriate columns to add is still\nan open problem and previous machine-learning-based approaches for CG only add\na constant quantity of columns per iteration due to the state-space explosion\nproblem. To address this, we propose Fast Family Column Generation (FFCG) -- a\nnovel reinforcement-learning-based CG that selects a variable number of columns\nas needed in an iteration. Specifically, we formulate the column selection\nproblem in CG as an MDP and design a reward metric that balances both the\nconvergence speed and the number of redundant columns. In our experiments, FFCG\nconverges faster on the common benchmarks and reduces the number of CG\niterations by 77.1% for Cutting Stock Problem (CSP) and 84.8% for Vehicle\nRouting Problem with Time Windows (VRPTW), and a 71.4% reduction in computing\ntime for CSP and 84.0% for VRPTW on average compared to several\nstate-of-the-art baselines.\n", "date": "2024-12-26", "categories": []} {"id": "2209.04265", "title": "Route Planning for Last-Mile Deliveries Using Mobile Parcel Lockers: A\n Hybrid Q-Learning Network Approach", "abstract": " Mobile parcel lockers have been recently proposed by logistics operators as a\ntechnology that could help reduce traffic congestion and operational costs in\nurban freight distribution. Given their ability to relocate throughout their\narea of deployment, they hold the potential to improve customer accessibility\nand convenience. In this study, we formulate the Mobile Parcel Locker Problem\n(MPLP) , a special case of the Location-Routing Problem (LRP) which determines\nthe optimal stopover location for MPLs throughout the day and plans\ncorresponding delivery routes. A Hybrid Q Learning Network based Method (HQM)\nis developed to resolve the computational complexity of the resulting large\nproblem instances while escaping local optima. In addition, the HQM is\nintegrated with global and local search mechanisms to resolve the dilemma of\nexploration and exploitation faced by classic reinforcement learning methods.\nWe examine the performance of HQM under different problem sizes (up to 200\nnodes) and benchmarked it against the exact approach and Genetic Algorithm\n(GA). Our results indicate that HQM achieves better optimisation performance\nwith shorter computation time than the exact approach solved by the Gurobi\nsolver in large problem instances. Additionally, the average reward obtained by\nHQM is 1.96 times greater than GA, which demonstrates that HQM has a better\noptimisation ability. Further, we identify critical factors that contribute to\nfleet size requirements, travel distances, and service delays. Our findings\noutline that the efficiency of MPLs is mainly contingent on the length of time\nwindows and the deployment of MPL stopovers. Finally, we highlight managerial\nimplications based on parametric analysis to provide guidance for logistics\noperators in the context of efficient last-mile distribution operations.\n", "date": "2022-09-09", "categories": []} {"id": "2412.17623", "title": "Towards An Unsupervised Learning Scheme for Efficiently Solving\n Parameterized Mixed-Integer Programs", "abstract": " In this paper, we describe a novel unsupervised learning scheme for\naccelerating the solution of a family of mixed integer programming (MIP)\nproblems. Distinct substantially from existing learning-to-optimize methods,\nour proposal seeks to train an autoencoder (AE) for binary variables in an\nunsupervised learning fashion, using data of optimal solutions to historical\ninstances for a parametric family of MIPs. By a deliberate design of AE\narchitecture and exploitation of its statistical implication, we present a\nsimple and straightforward strategy to construct a class of cutting plane\nconstraints from the decoder parameters of an offline-trained AE. These\nconstraints reliably enclose the optimal binary solutions of new problem\ninstances thanks to the representation strength of the AE. More importantly,\ntheir integration into the primal MIP problem leads to a tightened MIP with the\nreduced feasible region, which can be resolved at decision time using\noff-the-shelf solvers with much higher efficiency. Our method is applied to a\nbenchmark batch process scheduling problem formulated as a mixed integer linear\nprogramming (MILP) problem. Comprehensive results demonstrate that our approach\nsignificantly reduces the computational cost of off-the-shelf MILP solvers\nwhile retaining a high solution quality. The codes of this work are\nopen-sourced at https://github.com/qushiyuan/AE4BV.\n", "date": "2024-12-23", "categories": []} {"id": "2412.18571", "title": "Scalable Quantum-Inspired Optimization through Dynamic Qubit Compression", "abstract": " Hard combinatorial optimization problems, often mapped to Ising models,\npromise potential solutions with quantum advantage but are constrained by\nlimited qubit counts in near-term devices. We present an innovative\nquantum-inspired framework that dynamically compresses large Ising models to\nfit available quantum hardware of different sizes. Thus, we aim to bridge the\ngap between large-scale optimization and current hardware capabilities. Our\nmethod leverages a physics-inspired GNN architecture to capture complex\ninteractions in Ising models and accurately predict alignments among\nneighboring spins (aka qubits) at ground states. By progressively merging such\naligned spins, we can reduce the model size while preserving the underlying\noptimization structure. It also provides a natural trade-off between the\nsolution quality and size reduction, meeting different hardware constraints of\nquantum computing devices. Extensive numerical studies on Ising instances of\ndiverse topologies show that our method can reduce instance size at multiple\nlevels with virtually no losses in solution quality on the latest D-wave\nquantum annealers.\n", "date": "2024-12-24", "categories": []} {"id": "2412.11678", "title": "Loosely Synchronized Rule-Based Planning for Multi-Agent Path Finding\n with Asynchronous Actions", "abstract": " Multi-Agent Path Finding (MAPF) seeks collision-free paths for multiple\nagents from their respective starting locations to their respective goal\nlocations while minimizing path costs. Although many MAPF algorithms were\ndeveloped and can handle up to thousands of agents, they usually rely on the\nassumption that each action of the agent takes a time unit, and the actions of\nall agents are synchronized in a sense that the actions of agents start at the\nsame discrete time step, which may limit their use in practice. Only a few\nalgorithms were developed to address asynchronous actions, and they all lie on\none end of the spectrum, focusing on finding optimal solutions with limited\nscalability. This paper develops new planners that lie on the other end of the\nspectrum, trading off solution quality for scalability, by finding an unbounded\nsub-optimal solution for many agents. Our method leverages both search methods\n(LSS) in handling asynchronous actions and rule-based planning methods (PIBT)\nfor MAPF. We analyze the properties of our method and test it against several\nbaselines with up to 1000 agents in various maps. Given a runtime limit, our\nmethod can handle an order of magnitude more agents than the baselines with\nabout 25% longer makespan.\n", "date": "2024-12-16", "categories": []} {"id": "2412.15537", "title": "Enhancing Large-scale UAV Route Planing with Global and Local Features\n via Reinforcement Graph Fusion", "abstract": " Numerous remarkable advancements have been made in accuracy, speed, and\nparallelism for solving the Unmanned Aerial Vehicle Route Planing (UAVRP).\nHowever, existing UAVRP solvers face challenges when attempting to scale\neffectively and efficiently for larger instances. In this paper, we present a\ngeneralization framework that enables current UAVRP solvers to robustly extend\ntheir capabilities to larger instances, accommodating up to 10,000 points,\nusing widely recognized test sets. The UAVRP under a large number of patrol\npoints is a typical large-scale TSP problem.Our proposed framework comprises\nthree distinct steps. Firstly, we employ Delaunay triangulation to extract\nsubgraphs from large instances while preserving global features. Secondly, we\nutilize an embedded TSP solver to obtain sub-results, followed by graph fusion.\nFinally, we implement a decoding strategy customizable to the user's\nrequirements, resulting in high-quality solutions, complemented by a warming-up\nprocess for the heatmap. To demonstrate the flexibility of our approach, we\nintegrate two representative TSP solvers into our framework and conduct a\ncomprehensive comparative analysis against existing algorithms using large TSP\nbenchmark datasets. The results unequivocally demonstrate that our framework\nefficiently scales existing TSP solvers to handle large instances and\nconsistently outperforms state-of-the-art (SOTA) methods. Furthermore, since\nour proposed framework does not necessitate additional training or fine-tuning,\nwe believe that its generality can significantly advance research on end-to-end\nUAVRP solvers, enabling the application of a broader range of methods to\nreal-world scenarios.\n", "date": "2024-12-20", "categories": []} {"id": "2412.14995", "title": "HSEvo: Elevating Automatic Heuristic Design with Diversity-Driven\n Harmony Search and Genetic Algorithm Using LLMs", "abstract": " Automatic Heuristic Design (AHD) is an active research area due to its\nutility in solving complex search and NP-hard combinatorial optimization\nproblems in the real world. The recent advancements in Large Language Models\n(LLMs) introduce new possibilities by coupling LLMs with evolutionary\ncomputation to automatically generate heuristics, known as LLM-based\nEvolutionary Program Search (LLM-EPS). While previous LLM-EPS studies obtained\ngreat performance on various tasks, there is still a gap in understanding the\nproperties of heuristic search spaces and achieving a balance between\nexploration and exploitation, which is a critical factor in large heuristic\nsearch spaces. In this study, we address this gap by proposing two diversity\nmeasurement metrics and perform an analysis on previous LLM-EPS approaches,\nincluding FunSearch, EoH, and ReEvo. Results on black-box AHD problems reveal\nthat while EoH demonstrates higher diversity than FunSearch and ReEvo, its\nobjective score is unstable. Conversely, ReEvo's reflection mechanism yields\ngood objective scores but fails to optimize diversity effectively. With this\nfinding in mind, we introduce HSEvo, an adaptive LLM-EPS framework that\nmaintains a balance between diversity and convergence with a harmony search\nalgorithm. Through experimentation, we find that HSEvo achieved high diversity\nindices and good objective scores while remaining cost-effective. These results\nunderscore the importance of balancing exploration and exploitation and\nunderstanding heuristic search spaces in designing frameworks in LLM-EPS.\n", "date": "2024-12-19", "categories": []} {"id": "2401.08669", "title": "Deep Reinforcement Learning for Multi-Truck Vehicle Routing Problems\n with Multi-Leg Demand Routes", "abstract": " Deep reinforcement learning (RL) has been shown to be effective in producing\napproximate solutions to some vehicle routing problems (VRPs), especially when\nusing policies generated by encoder-decoder attention mechanisms. While these\ntechniques have been quite successful for relatively simple problem instances,\nthere are still under-researched and highly complex VRP variants for which no\neffective RL method has been demonstrated. In this work we focus on one such\nVRP variant, which contains multiple trucks and multi-leg routing requirements.\nIn these problems, demand is required to move along sequences of nodes, instead\nof just from a start node to an end node. With the goal of making deep RL a\nviable strategy for real-world industrial-scale supply chain logistics, we\ndevelop new extensions to existing encoder-decoder attention models which allow\nthem to handle multiple trucks and multi-leg routing requirements. Our models\nhave the advantage that they can be trained for a small number of trucks and\nnodes, and then embedded into a large supply chain to yield solutions for\nlarger numbers of trucks and nodes. We test our approach on a real supply chain\nenvironment arising in the operations of Japanese automotive parts manufacturer\nAisin Corporation, and find that our algorithm outperforms Aisin's previous\nbest solution.\n", "date": "2024-01-08", "categories": []} {"id": "2412.14052", "title": "Neural Combinatorial Optimization for Stochastic Flexible Job Shop\n Scheduling Problems", "abstract": " Neural combinatorial optimization (NCO) has gained significant attention due\nto the potential of deep learning to efficiently solve combinatorial\noptimization problems. NCO has been widely applied to job shop scheduling\nproblems (JSPs) with the current focus predominantly on deterministic problems.\nIn this paper, we propose a novel attention-based scenario processing module\n(SPM) to extend NCO methods for solving stochastic JSPs. Our approach\nexplicitly incorporates stochastic information by an attention mechanism that\ncaptures the embedding of sampled scenarios (i.e., an approximation of\nstochasticity). Fed with the embedding, the base neural network is intervened\nby the attended scenarios, which accordingly learns an effective policy under\nstochasticity. We also propose a training paradigm that works harmoniously with\neither the expected makespan or Value-at-Risk objective. Results demonstrate\nthat our approach outperforms existing learning and non-learning methods for\nthe flexible JSP problem with stochastic processing times on a variety of\ninstances. In addition, our approach holds significant generalizability to\nvaried numbers of scenarios and disparate distributions.\n", "date": "2024-12-18", "categories": []} {"id": "2408.02960", "title": "Anytime Multi-Agent Path Finding with an Adaptive Delay-Based Heuristic", "abstract": " Anytime multi-agent path finding (MAPF) is a promising approach to scalable\npath optimization in multi-agent systems. MAPF-LNS, based on Large Neighborhood\nSearch (LNS), is the current state-of-the-art approach where a fast initial\nsolution is iteratively optimized by destroying and repairing selected paths of\nthe solution. Current MAPF-LNS variants commonly use an adaptive selection\nmechanism to choose among multiple destroy heuristics. However, to determine\npromising destroy heuristics, MAPF-LNS requires a considerable amount of\nexploration time. As common destroy heuristics are non-adaptive, any\nperformance bottleneck caused by these heuristics cannot be overcome via\nadaptive heuristic selection alone, thus limiting the overall effectiveness of\nMAPF-LNS in terms of solution cost. In this paper, we propose Adaptive\nDelay-based Destroy-and-Repair Enhanced with Success-based Self-Learning\n(ADDRESS) as a single-destroy-heuristic variant of MAPF-LNS. ADDRESS applies\nrestricted Thompson Sampling to the top-K set of the most delayed agents to\nselect a seed agent for adaptive LNS neighborhood generation. We evaluate\nADDRESS in multiple maps from the MAPF benchmark set and demonstrate cost\nimprovements by at least 50% in large-scale scenarios with up to a thousand\nagents, compared with the original MAPF-LNS and other state-of-the-art methods.\n", "date": "2024-08-06", "categories": []} {"id": "2412.09444", "title": "Search Strategy Generation for Branch and Bound Using Genetic\n Programming", "abstract": " Branch-and-Bound (B\\&B) is an exact method in integer programming that\nrecursively divides the search space into a tree. During the resolution\nprocess, determining the next subproblem to explore within the tree-known as\nthe search strategy-is crucial. Hand-crafted heuristics are commonly used, but\nnone are effective over all problem classes. Recent approaches utilizing neural\nnetworks claim to make more intelligent decisions but are computationally\nexpensive. In this paper, we introduce GP2S (Genetic Programming for Search\nStrategy), a novel machine learning approach that automatically generates a\nB\\&B search strategy heuristic, aiming to make intelligent decisions while\nbeing computationally lightweight. We define a policy as a function that\nevaluates the quality of a B\\&B node by combining features from the node and\nthe problem; the search strategy policy is then defined by a best-first search\nbased on this node ranking. The policy space is explored using a genetic\nprogramming algorithm, and the policy that achieves the best performance on a\ntraining set is selected. We compare our approach with the standard method of\nthe SCIP solver, a recent graph neural network-based method, and handcrafted\nheuristics. Our first evaluation includes three types of primal hard problems,\ntested on instances similar to the training set and on larger instances. Our\nmethod is at most 2\\% slower than the best baseline and consistently\noutperforms SCIP, achieving an average speedup of 11.3\\%. Additionally, GP2S is\ntested on the MIPLIB 2017 dataset, generating multiple heuristics from\ndifferent subsets of instances. It exceeds SCIP's average performance in 7 out\nof 10 cases across 15 times more instances and under a time limit 15 times\nlonger, with some GP2S methods leading on most experiments in terms of the\nnumber of feasible solutions or optimality gap.\n", "date": "2024-12-12", "categories": []} {"id": "2408.12214", "title": "Bridging Large Language Models and Optimization: A Unified Framework for\n Text-attributed Combinatorial Optimization", "abstract": " To advance capabilities of large language models (LLMs) in solving\ncombinatorial optimization problems (COPs), this paper presents the\nLanguage-based Neural COP Solver (LNCS), a novel framework that is unified for\nthe end-to-end resolution of diverse text-attributed COPs. LNCS leverages LLMs\nto encode problem instances into a unified semantic space, and integrates their\nembeddings with a Transformer-based solution generator to produce high-quality\nsolutions. By training the solution generator with conflict-free multi-task\nreinforcement learning, LNCS effectively enhances LLM performance in tackling\nCOPs of varying types and sizes, achieving state-of-the-art results across\ndiverse problems. Extensive experiments validate the effectiveness and\ngeneralizability of the LNCS, highlighting its potential as a unified and\npractical framework for real-world COP applications.\n", "date": "2024-08-22", "categories": []} {"id": "2412.10682", "title": "Stochastic $k$-Submodular Bandits with Full Bandit Feedback", "abstract": " In this paper, we present the first sublinear $\\alpha$-regret bounds for\nonline $k$-submodular optimization problems with full-bandit feedback, where\n$\\alpha$ is a corresponding offline approximation ratio. Specifically, we\npropose online algorithms for multiple $k$-submodular stochastic combinatorial\nmulti-armed bandit problems, including (i) monotone functions and individual\nsize constraints, (ii) monotone functions with matroid constraints, (iii)\nnon-monotone functions with matroid constraints, (iv) non-monotone functions\nwithout constraints, and (v) monotone functions without constraints. We\ntransform approximation algorithms for offline $k$-submodular maximization\nproblems into online algorithms through the offline-to-online framework\nproposed by Nie et al. (2023a). A key contribution of our work is analyzing the\nrobustness of the offline algorithms.\n", "date": "2024-12-14", "categories": []} {"id": "2412.10163", "title": "Scaling Combinatorial Optimization Neural Improvement Heuristics with\n Online Search and Adaptation", "abstract": " We introduce Limited Rollout Beam Search (LRBS), a beam search strategy for\ndeep reinforcement learning (DRL) based combinatorial optimization improvement\nheuristics. Utilizing pre-trained models on the Euclidean Traveling Salesperson\nProblem, LRBS significantly enhances both in-distribution performance and\ngeneralization to larger problem instances, achieving optimality gaps that\noutperform existing improvement heuristics and narrowing the gap with\nstate-of-the-art constructive methods. We also extend our analysis to two\npickup and delivery TSP variants to validate our results. Finally, we employ\nour search strategy for offline and online adaptation of the pre-trained\nimprovement policy, leading to improved search performance and surpassing\nrecent adaptive methods for constructive heuristics.\n", "date": "2024-12-13", "categories": []} {"id": "2412.09090", "title": "Integrated trucks assignment and scheduling problem with mixed service\n mode docks: A Q-learning based adaptive large neighborhood search algorithm", "abstract": " Mixed service mode docks enhance efficiency by flexibly handling both loading\nand unloading trucks in warehouses. However, existing research often\npredetermines the number and location of these docks prior to planning truck\nassignment and sequencing. This paper proposes a new model integrating dock\nmode decision, truck assignment, and scheduling, thus enabling adaptive dock\nmode arrangements. Specifically, we introduce a Q-learning-based adaptive large\nneighborhood search (Q-ALNS) algorithm to address the integrated problem. The\nalgorithm adjusts dock modes via perturbation operators, while truck assignment\nand scheduling are solved using destroy and repair local search operators.\nQ-learning adaptively selects these operators based on their performance\nhistory and future gains, employing the epsilon-greedy strategy. Extensive\nexperimental results and statistical analysis indicate that the Q-ALNS benefits\nfrom efficient operator combinations and its adaptive mechanism, consistently\noutperforming benchmark algorithms in terms of optimality gap and Pareto front\ndiscovery. In comparison to the predetermined service mode, our adaptive\nstrategy results in lower average tardiness and makespan, highlighting its\nsuperior adaptability to varying demands.\n", "date": "2024-12-12", "categories": []} {"id": "2412.08287", "title": "DistrictNet: Decision-aware learning for geographical districting", "abstract": " Districting is a complex combinatorial problem that consists in partitioning\na geographical area into small districts. In logistics, it is a major strategic\ndecision determining operating costs for several years. Solving districting\nproblems using traditional methods is intractable even for small geographical\nareas and existing heuristics often provide sub-optimal results. We present a\nstructured learning approach to find high-quality solutions to real-world\ndistricting problems in a few minutes. It is based on integrating a\ncombinatorial optimization layer, the capacitated minimum spanning tree\nproblem, into a graph neural network architecture. To train this pipeline in a\ndecision-aware fashion, we show how to construct target solutions embedded in a\nsuitable space and learn from target solutions. Experiments show that our\napproach outperforms existing methods as it can significantly reduce costs on\nreal-world cities.\n", "date": "2024-12-11", "categories": []} {"id": "2406.14096", "title": "Graph Neural Networks for Job Shop Scheduling Problems: A Survey", "abstract": " Job shop scheduling problems (JSSPs) represent a critical and challenging\nclass of combinatorial optimization problems. Recent years have witnessed a\nrapid increase in the application of graph neural networks (GNNs) to solve\nJSSPs, albeit lacking a systematic survey of the relevant literature. This\npaper aims to thoroughly review prevailing GNN methods for different types of\nJSSPs and the closely related flow-shop scheduling problems (FSPs), especially\nthose leveraging deep reinforcement learning (DRL). We begin by presenting the\ngraph representations of various JSSPs, followed by an introduction to the most\ncommonly used GNN architectures. We then review current GNN-based methods for\neach problem type, highlighting key technical elements such as graph\nrepresentations, GNN architectures, GNN tasks, and training algorithms.\nFinally, we summarize and analyze the advantages and limitations of GNNs in\nsolving JSSPs and provide potential future research opportunities. We hope this\nsurvey can motivate and inspire innovative approaches for more powerful\nGNN-based approaches in tackling JSSPs and other scheduling problems.\n", "date": "2024-06-20", "categories": []} {"id": "2403.11351", "title": "A Semidefinite Programming-Based Branch-and-Cut Algorithm for\n Biclustering", "abstract": " Biclustering, also called co-clustering, block clustering, or two-way\nclustering, involves the simultaneous clustering of both the rows and columns\nof a data matrix into distinct groups, such that the rows and columns within a\ngroup display similar patterns. As a model problem for biclustering, we\nconsider the $k$-densest-disjoint biclique problem, whose goal is to identify\n$k$ disjoint complete bipartite subgraphs (called bicliques) of a given\nweighted complete bipartite graph such that the sum of their densities is\nmaximized. To address this problem, we present a tailored branch-and-cut\nalgorithm. For the upper bound routine, we consider a semidefinite programming\nrelaxation and propose valid inequalities to strengthen the bound. We solve\nthis relaxation in a cutting-plane fashion using a first-order method. For the\nlower bound, we design a maximum weight matching rounding procedure that\nexploits the solution of the relaxation solved at each node. Computational\nresults on both synthetic and real-world instances show that the proposed\nalgorithm can solve instances approximately 20 times larger than those handled\nby general-purpose solvers.\n", "date": "2024-03-17", "categories": []} {"id": "2412.04256", "title": "Transient Multi-Agent Path Finding for Lifelong Navigation in Dense\n Environments", "abstract": " Multi-Agent Path Finding (MAPF) deals with finding conflict-free paths for a\nset of agents from an initial configuration to a given target configuration.\nThe Lifelong MAPF (LMAPF) problem is a well-studied online version of MAPF in\nwhich an agent receives a new target when it reaches its current target. The\ncommon approach for solving LMAPF is to treat it as a sequence of MAPF\nproblems, periodically replanning from the agents' current configurations to\ntheir current targets. A significant drawback in this approach is that in MAPF\nthe agents must reach a configuration in which all agents are at their targets\nsimultaneously, which is needlessly restrictive for LMAPF. Techniques have been\nproposed to indirectly mitigate this drawback. We describe cases where these\nmitigation techniques fail. As an alternative, we propose to solve LMAPF\nproblems by solving a sequence of modified MAPF problems, in which the\nobjective is for each agent to eventually visit its target, but not necessarily\nfor all agents to do so simultaneously. We refer to this MAPF variant as\nTransient MAPF (TMAPF) and propose several algorithms for solving it based on\nexisting MAPF algorithms. A limited experimental evaluation identifies some\ncases where using a TMAPF algorithm instead of a MAPF algorithm with an LMAPF\nframework can improve the system throughput significantly.\n", "date": "2024-12-05", "categories": []} {"id": "2410.07980", "title": "D-Wave's Nonlinear-Program Hybrid Solver: Description and Performance\n Analysis", "abstract": " The development of advanced quantum-classical algorithms is among the most\nprominent strategies in quantum computing. Numerous hybrid solvers have been\nintroduced recently. Many of these methods are created ad hoc to address\nspecific use cases. However, several well-established schemes are frequently\nutilized to address optimization problems. In this context, D-Wave launched the\nHybrid Solver Service in 2020, offering a portfolio of methods designed to\naccelerate time-to-solution for users aiming to optimize performance and\noperational processes. Recently, a new technique has been added to this\nportfolio: the Nonlinear-Program Hybrid Solver. This paper describes this\nsolver and evaluates its performance through a benchmark of 45 instances across\nthree combinatorial optimization problems: the Traveling Salesman Problem, the\nKnapsack Problem, and the Maximum Cut Problem. To facilitate the use of this\nrelatively unexplored solver, we provide details of the implementation used to\nsolve these three optimization problems.\n", "date": "2024-10-10", "categories": []} {"id": "2308.10656", "title": "Practical Parallel Algorithms for Non-Monotone Submodular Maximization", "abstract": " Submodular maximization has found extensive applications in various domains\nwithin the field of artificial intelligence, including but not limited to\nmachine learning, computer vision, and natural language processing. With the\nincreasing size of datasets in these domains, there is a pressing need to\ndevelop efficient and parallelizable algorithms for submodular maximization.\nOne measure of the parallelizability of a submodular maximization algorithm is\nits adaptive complexity, which indicates the number of sequential rounds where\na polynomial number of queries to the objective function can be executed in\nparallel. In this paper, we study the problem of non-monotone submodular\nmaximization subject to a knapsack constraint, and propose the first\ncombinatorial algorithm achieving an $(8+\\epsilon)$-approximation under\n$\\mathcal{O}(\\log n)$ adaptive complexity, which is \\textit{optimal} up to a\nfactor of $\\mathcal{O}(\\log\\log n)$. Moreover, we also propose the first\nalgorithm with both provable approximation ratio and sublinear adaptive\ncomplexity for the problem of non-monotone submodular maximization subject to a\n$k$-system constraint. As a by-product, we show that our two algorithms can\nalso be applied to the special case of submodular maximization subject to a\ncardinality constraint, and achieve performance bounds comparable with those of\nstate-of-the-art algorithms. Finally, the effectiveness of our approach is\ndemonstrated by extensive experiments on real-world applications.\n", "date": "2023-08-21", "categories": []} {"id": "2412.01937", "title": "Approximately Optimal Search on a Higher-dimensional Sliding Puzzle", "abstract": " Higher-dimensional sliding puzzles are constructed on the vertices of a\n$d$-dimensional hypercube, where $2^d-l$ vertices are distinctly coloured.\nRings with the same colours are initially set randomly on the vertices of the\nhypercube. The goal of the puzzle is to move each of the $2^d-l$ rings to\npre-defined target vertices on the cube. In this setting, the $k$-rule\nconstraint represents a generalisation of edge collision for the movement of\ncolours between vertices, allowing movement only when a hypercube face of\ndimension $k$ containing a ring is completely free of other rings. Starting\nfrom an initial configuration, what is the minimum number of moves needed to\nmake ring colours match the vertex colours? An algorithm that provides us with\nsuch a number is called God's algorithm. When such an algorithm exists, it does\nnot have a polynomial time complexity, at least in the case of the 15-puzzle\ncorresponding to $k=1$ in the cubical puzzle. This paper presents a\ncomprehensive computational study of different scenarios of the\nhigher-dimensional puzzle. A benchmark of three computational techniques, an\nexact algorithm (the A* search) and two approximately optimal search techniques\n(an evolutionary algorithm (EA) and reinforcement learning (RL)) is presented\nin this work. The experiments show that all three methods can successfully\nsolve the puzzle of dimension three for different face dimensions and across\nvarious difficulty levels. When the dimension increases, the A* search fails,\nand RL and EA methods can still provide a generally acceptable solution, i.e. a\ndistribution of a number of moves with a median value of less than $30$.\nOverall, the EA method consistently requires less computational time, while\nfailing in most cases to minimise the number of moves for the puzzle dimensions\n$d=4$ and $d=5$.\n", "date": "2024-12-02", "categories": []} {"id": "2412.00346", "title": "CaDA: Cross-Problem Routing Solver with Constraint-Aware Dual-Attention", "abstract": " Vehicle Routing Problems (VRPs) are significant Combinatorial Optimization\n(CO) problems holding substantial practical importance. Recently, Neural\nCombinatorial Optimization (NCO), which involves training deep learning models\non extensive data to learn vehicle routing heuristics, has emerged as a\npromising approach due to its efficiency and the reduced need for manual\nalgorithm design. However, applying NCO across diverse real-world scenarios\nwith various constraints necessitates cross-problem capabilities. Current NCO\nmethods typically employ a unified model lacking a constraint-specific\nstructure, thereby restricting their cross-problem performance. Current\nmulti-task methods for VRPs typically employ a constraint-unaware model,\nlimiting their cross-problem performance. Furthermore, they rely solely on\nglobal connectivity, which fails to focus on key nodes and leads to inefficient\nrepresentation learning. This paper introduces a Constraint-Aware\nDual-Attention Model (CaDA), designed to address these limitations. CaDA\nincorporates a constraint prompt that efficiently represents different problem\nvariants. Additionally, it features a dual-attention mechanism with a global\nbranch for capturing broader graph-wide information and a sparse branch that\nselectively focuses on the most relevant nodes. We comprehensively evaluate our\nmodel on 16 different VRPs and compare its performance against existing\ncross-problem VRP solvers. CaDA achieves state-of-the-art results across all\nthe VRPs. Our ablation study further confirms that each component of CaDA\ncontributes positively to its cross-problem learning performance.\n", "date": "2024-11-30", "categories": []} {"id": "2412.01094", "title": "A Hierarchical Heuristic for Clustered Steiner Trees in the Plane with\n Obstacles", "abstract": " Euclidean Steiner trees are relevant to model minimal networks in real-world\napplications ubiquitously. In this paper, we study the feasibility of a\nhierarchical approach embedded with bundling operations to compute multiple and\nmutually disjoint Euclidean Steiner trees that avoid clutter and overlapping\nwith obstacles in the plane, which is significant to model the decentralized\nand the multipoint coordination of agents in constrained 2D domains. Our\ncomputational experiments using arbitrary obstacle configuration with convex\nand non-convex geometries show the feasibility and the attractive performance\nwhen computing multiple obstacle-avoiding Steiner trees in the plane. Our\nresults offer the mechanisms to elucidate new operators for obstacle-avoiding\nSteiner trees.\n", "date": "2024-12-02", "categories": []} {"id": "2411.19734", "title": "A Note on Small Percolating Sets on Hypercubes via Generative AI", "abstract": " We apply a generative AI pattern-recognition technique called PatternBoost to\nstudy bootstrap percolation on hypercubes. With this, we slightly improve the\nbest existing upper bound for the size of percolating subsets of the hypercube.\n", "date": "2024-11-29", "categories": []} {"id": "2411.19744", "title": "Amplifying human performance in combinatorial competitive programming", "abstract": " Recent years have seen a significant surge in complex AI systems for\ncompetitive programming, capable of performing at admirable levels against\nhuman competitors. While steady progress has been made, the highest percentiles\nstill remain out of reach for these methods on standard competition platforms\nsuch as Codeforces. Here we instead focus on combinatorial competitive\nprogramming, where the target is to find as-good-as-possible solutions to\notherwise computationally intractable problems, over specific given inputs. We\nhypothesise that this scenario offers a unique testbed for human-AI synergy, as\nhuman programmers can write a backbone of a heuristic solution, after which AI\ncan be used to optimise the scoring function used by the heuristic. We deploy\nour approach on previous iterations of Hash Code, a global team programming\ncompetition inspired by NP-hard software engineering problems at Google, and we\nleverage FunSearch to evolve our scoring functions. Our evolved solutions\nsignificantly improve the attained scores from their baseline, successfully\nbreaking into the top percentile on all previous Hash Code online qualification\nrounds, and outperforming the top human teams on several. Our method is also\nperformant on an optimisation problem that featured in a recent held-out\nAtCoder contest.\n", "date": "2024-11-29", "categories": []} {"id": "2411.18321", "title": "Learning optimal objective values for MILP", "abstract": " Modern Mixed Integer Linear Programming (MILP) solvers use the\nBranch-and-Bound algorithm together with a plethora of auxiliary components\nthat speed up the search. In recent years, there has been an explosive\ndevelopment in the use of machine learning for enhancing and supporting these\nalgorithmic components. Within this line, we propose a methodology for\npredicting the optimal objective value, or, equivalently, predicting if the\ncurrent incumbent is optimal. For this task, we introduce a predictor based on\na graph neural network (GNN) architecture, together with a set of dynamic\nfeatures. Experimental results on diverse benchmarks demonstrate the efficacy\nof our approach, achieving high accuracy in the prediction task and\noutperforming existing methods. These findings suggest new opportunities for\nintegrating ML-driven predictions into MILP solvers, enabling smarter\ndecision-making and improved performance.\n", "date": "2024-11-27", "categories": []} {"id": "2411.17353", "title": "Joint Combinatorial Node Selection and Resource Allocations in the\n Lightning Network using Attention-based Reinforcement Learning", "abstract": " The Lightning Network (LN) has emerged as a second-layer solution to\nBitcoin's scalability challenges. The rise of Payment Channel Networks (PCNs)\nand their specific mechanisms incentivize individuals to join the network for\nprofit-making opportunities. According to the latest statistics, the total\nvalue locked within the Lightning Network is approximately \\$500 million.\nMeanwhile, joining the LN with the profit-making incentives presents several\nobstacles, as it involves solving a complex combinatorial problem that\nencompasses both discrete and continuous control variables related to node\nselection and resource allocation, respectively. Current research inadequately\ncaptures the critical role of resource allocation and lacks realistic\nsimulations of the LN routing mechanism. In this paper, we propose a Deep\nReinforcement Learning (DRL) framework, enhanced by the power of transformers,\nto address the Joint Combinatorial Node Selection and Resource Allocation\n(JCNSRA) problem. We have improved upon an existing environment by introducing\nmodules that enhance its routing mechanism, thereby narrowing the gap with the\nactual LN routing system and ensuring compatibility with the JCNSRA problem. We\ncompare our model against several baselines and heuristics, demonstrating its\nsuperior performance across various settings. Additionally, we address concerns\nregarding centralization in the LN by deploying our agent within the network\nand monitoring the centrality measures of the evolved graph. Our findings\nsuggest not only an absence of conflict between LN's decentralization goals and\nindividuals' revenue-maximization incentives but also a positive association\nbetween the two.\n", "date": "2024-11-26", "categories": []} {"id": "2405.20543", "title": "Towards a General Recipe for Combinatorial Optimization with\n Multi-Filter GNNs", "abstract": " Graph neural networks (GNNs) have achieved great success for a variety of\ntasks such as node classification, graph classification, and link prediction.\nHowever, the use of GNNs (and machine learning more generally) to solve\ncombinatorial optimization (CO) problems is much less explored. Here, we\nintroduce GCON, a novel GNN architecture that leverages a complex filter bank\nand localized attention mechanisms to solve CO problems on graphs. We show how\nour method differentiates itself from prior GNN-based CO solvers and how it can\nbe effectively applied to the maximum cut, minimum dominating set, and maximum\nclique problems in a unsupervised learning setting. GCON is competitive across\nall tasks and consistently outperforms other specialized GNN-based approaches,\nand is on par with the powerful Gurobi solver on the max-cut problem. We\nprovide an open-source implementation of our work at\nhttps://github.com/WenkelF/copt.\n", "date": "2024-05-31", "categories": []} {"id": "2311.07633", "title": "Benchmarking PtO and PnO Methods in the Predictive Combinatorial\n Optimization Regime", "abstract": " Predictive combinatorial optimization, where the parameters of combinatorial\noptimization (CO) are unknown at the decision-making time, is the precise\nmodeling of many real-world applications, including energy cost-aware\nscheduling and budget allocation on advertising. Tackling such a problem\nusually involves a prediction model and a CO solver. These two modules are\nintegrated into the predictive CO pipeline following two design principles:\n\"Predict-then-Optimize (PtO)\", which learns predictions by supervised training\nand subsequently solves CO using predicted coefficients, while the other, named\n\"Predict-and-Optimize (PnO)\", directly optimizes towards the ultimate decision\nquality and claims to yield better decisions than traditional PtO approaches.\nHowever, there lacks a systematic benchmark of both approaches, including the\nspecific design choices at the module level, as well as an evaluation dataset\nthat covers representative real-world scenarios. To this end, we develop a\nmodular framework to benchmark 11 existing PtO/PnO methods on 8 problems,\nincluding a new industrial dataset for combinatorial advertising that will be\nreleased. Our study shows that PnO approaches are better than PtO on 7 out of 8\nbenchmarks, but there is no silver bullet found for the specific design choices\nof PnO. A comprehensive categorization of current approaches and integration of\ntypical scenarios are provided under a unified benchmark. Therefore, this paper\ncould serve as a comprehensive benchmark for future PnO approach development\nand also offer fast prototyping for application-focused development. The code\nis available at https://github.com/Thinklab-SJTU/PredictiveCO-Benchmark.\n", "date": "2023-11-13", "categories": []} {"id": "2403.20212", "title": "On Size and Hardness Generalization in Unsupervised Learning for the\n Travelling Salesman Problem", "abstract": " We study the generalization capability of Unsupervised Learning in solving\nthe Travelling Salesman Problem (TSP). We use a Graph Neural Network (GNN)\ntrained with a surrogate loss function to generate an embedding for each node.\nWe use these embeddings to construct a heat map that indicates the likelihood\nof each edge being part of the optimal route. We then apply local search to\ngenerate our final predictions. Our investigation explores how different\ntraining instance sizes, embedding dimensions, and distributions influence the\noutcomes of Unsupervised Learning methods. Our results show that training with\nlarger instance sizes and increasing embedding dimensions can build a more\neffective representation, enhancing the model's ability to solve TSP.\nFurthermore, in evaluating generalization across different distributions, we\nfirst determine the hardness of various distributions and explore how different\nhardnesses affect the final results. Our findings suggest that models trained\non harder instances exhibit better generalization capabilities, highlighting\nthe importance of selecting appropriate training instances in solving TSP using\nUnsupervised Learning.\n", "date": "2024-03-29", "categories": []} {"id": "2411.05263", "title": "Minimal Conditions for Beneficial Neighbourhood Search and Local Descent", "abstract": " This paper investigates what properties a neighbourhood requires to support\nbeneficial local search. We show that neighbourhood locality, and a reduction\nin cost probability towards the optimum, support a proof that search among\nneighbours is more likely to find an improving solution in a single search step\nthan blind search. This is the first paper to introduce such a proof. The\nconcepts underlying these properties are illustrated on a satisfiability\nproblem class, and on travelling salesman problems. Secondly, for a given cost\ntarget t, we investigate a combination of blind search and local descent termed\nlocal blind descent, and present various conditions under which the expected\nnumber of steps to reach a cost better than t using local blind descent, is\nproven to be smaller than with blind search. Experiments indicate that local\nblind descent, given target cost t, should switch to local descent at a\nstarting cost that reduces as t approaches the optimum.\n", "date": "2024-11-08", "categories": []} {"id": "2210.01021", "title": "Budget-Aware Sequential Brick Assembly with Efficient Constraint\n Satisfaction", "abstract": " We tackle the problem of sequential brick assembly with LEGO bricks to create\ncombinatorial 3D structures. This problem is challenging since this brick\nassembly task encompasses the characteristics of combinatorial optimization\nproblems. In particular, the number of assemblable structures increases\nexponentially as the number of bricks used increases. To solve this problem, we\npropose a new method to predict the scores of the next brick position by\nemploying a U-shaped sparse 3D convolutional neural network. Along with the 3D\nconvolutional network, a one-initialized brick-sized convolution filter is used\nto efficiently validate assembly constraints between bricks without training\nitself. By the nature of this one-initialized convolution filter, we can\nreadily consider several different brick types by benefiting from modern\nimplementation of convolution operations. To generate a novel structure, we\ndevise a sampling strategy to determine the next brick position considering the\nsatisfaction of assembly constraints. Moreover, our method is designed for\neither budget-free or budget-aware scenario where a budget may confine the\nnumber of bricks and their types. We demonstrate that our method successfully\ngenerates a variety of brick structures and outperforms existing methods with\nBayesian optimization, deep graph generative model, and reinforcement learning.\n", "date": "2022-10-03", "categories": []} {"id": "2411.09068", "title": "Liner Shipping Network Design with Reinforcement Learning", "abstract": " This paper proposes a novel reinforcement learning framework to address the\nLiner Shipping Network Design Problem (LSNDP), a challenging combinatorial\noptimization problem focused on designing cost-efficient maritime shipping\nroutes. Traditional methods for solving the LSNDP typically involve decomposing\nthe problem into sub-problems, such as network design and multi-commodity flow,\nwhich are then tackled using approximate heuristics or large neighborhood\nsearch (LNS) techniques. In contrast, our approach employs a model-free\nreinforcement learning algorithm on the network design, integrated with a\nheuristic-based multi-commodity flow solver, to produce competitive results on\nthe publicly available LINERLIB benchmark. Additionally, our method also\ndemonstrates generalization capabilities by producing competitive solutions on\nthe benchmark instances after training on perturbed instances.\n", "date": "2024-11-13", "categories": []} {"id": "2411.05798", "title": "A Genetic Algorithm for Multi-Capacity Fixed-Charge Flow Network Design", "abstract": " The Multi-Capacity Fixed-Charge Network Flow (MC-FCNF) problem, a\ngeneralization of the Fixed-Charge Network Flow problem, aims to assign\ncapacities to edges in a flow network such that a target amount of flow can be\nhosted at minimum cost. The cost model for both problems dictates that the\nfixed cost of an edge is incurred for any non-zero amount of flow hosted by\nthat edge. This problem naturally arises in many areas including infrastructure\ndesign, transportation, telecommunications, and supply chain management. The\nMC-FCNF problem is NP-Hard, so solving large instances using exact techniques\nis impractical. This paper presents a genetic algorithm designed to quickly\nfind high-quality flow solutions to the MC-FCNF problem. The genetic algorithm\nuses a novel solution representation scheme that eliminates the need to repair\ninvalid flow solutions, which is an issue common to many other genetic\nalgorithms for the MC-FCNF problem. The genetic algorithm's efficiency is\ndisplayed with an evaluation using real-world CO2 capture and storage\ninfrastructure design data. The evaluation results highlight the genetic\nalgorithm's potential for solving large-scale network design problems.\n", "date": "2024-10-26", "categories": []} {"id": "2411.05834", "title": "Assessing and Enhancing Graph Neural Networks for Combinatorial\n Optimization: Novel Approaches and Application in Maximum Independent Set\n Problems", "abstract": " Combinatorial optimization (CO) problems are challenging as the computation\ntime grows exponentially with the input. Graph Neural Networks (GNNs) show\npromise for researchers in solving CO problems. This study investigates the\neffectiveness of GNNs in solving the maximum independent set (MIS) problem,\ninspired by the intriguing findings of Schuetz et al., and aimed to enhance\nthis solver. Despite the promise shown by GNNs, some researchers observed\ndiscrepancies when reproducing the findings, particularly compared to the\ngreedy algorithm, for instance. We reproduced Schuetz' Quadratic Unconstrained\nBinary Optimization (QUBO) unsupervised approach and explored the possibility\nof combining it with a supervised learning approach for solving MIS problems.\nWhile the QUBO unsupervised approach did not guarantee maximal or optimal\nsolutions, it provided a solid first guess for post-processing techniques like\ngreedy decoding or tree-based methods. Moreover, our findings indicated that\nthe supervised approach could further refine the QUBO unsupervised solver, as\nthe learned model assigned meaningful probabilities for each node as initial\nnode features, which could then be improved with the QUBO unsupervised\napproach. Thus, GNNs offer a valuable method for solving CO problems by\nintegrating learned graph structures rather than relying solely on traditional\nheuristic functions. This research highlights the potential of GNNs to boost\nsolver performance by leveraging ground truth during training and using\noptimization functions to learn structural graph information, marking a\npioneering step towards improving prediction accuracy in a non-autoregressive\nmanner.\n", "date": "2024-11-06", "categories": []} {"id": "2411.05318", "title": "Fairness in Monotone $k$-submodular Maximization: Algorithms and\n Applications", "abstract": " Submodular optimization has become increasingly prominent in machine learning\nand fairness has drawn much attention. In this paper, we propose to study the\nfair $k$-submodular maximization problem and develop a\n$\\frac{1}{3}$-approximation greedy algorithm with a running time of\n$\\mathcal{O}(knB)$. To the best of our knowledge, our work is the first to\nincorporate fairness in the context of $k$-submodular maximization, and our\ntheoretical guarantee matches the best-known $k$-submodular maximization\nresults without fairness constraints. In addition, we have developed a faster\nthreshold-based algorithm that achieves a $(\\frac{1}{3} - \\epsilon)$\napproximation with $\\mathcal{O}(\\frac{kn}{\\epsilon} \\log \\frac{B}{\\epsilon})$\nevaluations of the function $f$. Furthermore, for both algorithms, we provide\napproximation guarantees when the $k$-submodular function is not accessible but\nonly can be approximately accessed. We have extensively validated our\ntheoretical findings through empirical research and examined the practical\nimplications of fairness. Specifically, we have addressed the question: ``What\nis the price of fairness?\" through case studies on influence maximization with\n$k$ topics and sensor placement with $k$ types. The experimental results show\nthat the fairness constraints do not significantly undermine the quality of\nsolutions.\n", "date": "2024-11-08", "categories": []} {"id": "2405.18681", "title": "A random-key GRASP for combinatorial optimization", "abstract": " This paper proposes a problem-independent GRASP metaheuristic using the\nrandom-key optimizer (RKO) paradigm. GRASP (greedy randomized adaptive search\nprocedure) is a metaheuristic for combinatorial optimization that repeatedly\napplies a semi-greedy construction procedure followed by a local search\nprocedure. The best solution found over all iterations is returned as the\nsolution of the GRASP. Continuous GRASP (C-GRASP) is an extension of GRASP for\ncontinuous optimization in the unit hypercube. A random-key optimizer (RKO)\nuses a vector of random keys to encode a solution to a combinatorial\noptimization problem. It uses a decoder to evaluate a solution encoded by the\nvector of random keys. A random-key GRASP is a C-GRASP where points in the unit\nhypercube are evaluated employing a decoder. We describe random key GRASP\nconsisting of a problem-independent component and a problem-dependent decoder.\nAs a proof of concept, the random-key GRASP is tested on five NP-hard\ncombinatorial optimization problems: traveling salesman problem, tree of hubs\nlocation problem, Steiner triple covering problem, node capacitated graph\npartitioning problem, and job sequencing and tool switching problem.\n", "date": "2024-05-29", "categories": []} {"id": "2411.04777", "title": "Learn to Solve Vehicle Routing Problems ASAP: A Neural Optimization\n Approach for Time-Constrained Vehicle Routing Problems with Finite Vehicle\n Fleet", "abstract": " Finding a feasible and prompt solution to the Vehicle Routing Problem (VRP)\nis a prerequisite for efficient freight transportation, seamless logistics, and\nsustainable mobility. Traditional optimization methods reach their limits when\nconfronted with the real-world complexity of VRPs, which involve numerous\nconstraints and objectives. Recently, the ability of generative Artificial\nIntelligence (AI) to solve combinatorial tasks, known as Neural Combinatorial\nOptimization (NCO), demonstrated promising results, offering new perspectives.\nIn this study, we propose an NCO approach to solve a time-constrained\ncapacitated VRP with a finite vehicle fleet size. The approach is based on an\nencoder-decoder architecture, formulated in line with the Policy Optimization\nwith Multiple Optima (POMO) protocol and trained via a Proximal Policy\nOptimization (PPO) algorithm. We successfully trained the policy with multiple\nobjectives (minimizing the total distance while maximizing vehicle utilization)\nand evaluated it on medium and large instances, benchmarking it against\nstate-of-the-art heuristics. The method is able to find adequate and\ncost-efficient solutions, showing both flexibility and robust generalization.\nFinally, we provide a critical analysis of the solution generated by NCO and\ndiscuss the challenges and opportunities of this new branch of intelligent\nlearning algorithms emerging in optimization science, focusing on freight\ntransportation.\n", "date": "2024-11-07", "categories": []} {"id": "2404.08837", "title": "Vehicle-to-Vehicle Charging: Model, Complexity, and Heuristics", "abstract": " The rapid adoption of Electric Vehicles (EVs) poses challenges for\nelectricity grids to accommodate or mitigate peak demand. Vehicle-to-Vehicle\nCharging (V2VC) has been recently adopted by popular EVs, posing new\nopportunities and challenges to the management and operation of EVs. We present\na novel V2VC model that allows decision-makers to take V2VC into account when\noptimizing their EV operations. We show that optimizing V2VC is NP-Complete and\nfind that even small problem instances are computationally challenging. We\npropose R-V2VC, a heuristic that takes advantage of the resulting totally\nunimodular constraint matrix to efficiently solve problems of realistic sizes.\nOur results demonstrate that R-V2VC presents a linear growth in the solution\ntime as the problem size increases, while achieving solutions of optimal or\nnear-optimal quality. R-V2VC can be used for real-world operations and to study\nwhat-if scenarios when evaluating the costs and benefits of V2VC.\n", "date": "2024-04-12", "categories": []} {"id": "2411.03324", "title": "A Surrogate Model for Quay Crane Scheduling Problem", "abstract": " In ports, a variety of tasks are carried out, and scheduling these tasks is\ncrucial due to its significant impact on productivity, making the generation of\nprecise plans essential. This study proposes a method to solve the Quay Crane\nScheduling Problem (QCSP), a representative task scheduling problem in ports\nknown to be NP-Hard, more quickly and accurately. First, the study suggests a\nmethod to create more accurate work plans for Quay Cranes (QCs) by learning\nfrom actual port data to accurately predict the working speed of QCs. Next, a\nSurrogate Model is proposed by combining a Machine Learning (ML) model with a\nGenetic Algorithm (GA), which is widely used to solve complex optimization\nproblems, enabling faster and more precise exploration of solutions. Unlike\nmethods that use fixed-dimensional chromosome encoding, the proposed\nmethodology can provide solutions for encodings of various dimensions. To\nvalidate the performance of the newly proposed methodology, comparative\nexperiments were conducted, demonstrating faster search speeds and improved\nfitness scores. The method proposed in this study can be applied not only to\nQCSP but also to various NP-Hard problems, and it opens up possibilities for\nthe further development of advanced search algorithms by combining heuristic\nalgorithms with ML models.\n", "date": "2024-10-22", "categories": []} {"id": "2402.02774", "title": "Accelerating Matroid Optimization through Fast Imprecise Oracles", "abstract": " Querying complex models for precise information (e.g. traffic models,\ndatabase systems, large ML models) often entails intense computations and\nresults in long response times. Thus, weaker models which give imprecise\nresults quickly can be advantageous, provided inaccuracies can be resolved\nusing few queries to a stronger model. In the fundamental problem of computing\na maximum-weight basis of a matroid, a well-known generalization of many\ncombinatorial optimization problems, algorithms have access to a clean oracle\nto query matroid information. We additionally equip algorithms with a fast but\ndirty oracle modelling an unknown, potentially different matroid. We design and\nanalyze practical algorithms which only use few clean queries w.r.t. the\nquality of the dirty oracle, while maintaining robustness against arbitrarily\npoor dirty matroids, approaching the performance of classic algorithms for the\ngiven problem. Notably, we prove that our algorithms are, in many respects,\nbest-possible. Further, we outline extensions to other matroid oracle types,\nnon-free dirty oracles and other matroid problems.\n", "date": "2024-02-05", "categories": []} {"id": "2310.04345", "title": "Deep Learning for Two-Stage Robust Integer Optimization", "abstract": " Robust optimization is an established framework for modeling optimization\nproblems with uncertain parameters. While static robust optimization is often\ncriticized for being too conservative, two-stage (or adjustable) robust\noptimization (2RO) provides a less conservative alternative by allowing some\ndecisions to be made after the uncertain parameters have been revealed.\nUnfortunately, in the case of integer decision variables, existing solution\nmethods for 2RO typically fail to solve large-scale instances, limiting the\napplicability of this modeling paradigm to simple cases. We propose Neur2RO, a\ndeep-learning-augmented instantiation of the column-and-constraint-generation\n(CCG) algorithm, which expands the applicability of the 2RO framework to\nlarge-scale instances with integer decisions in both stages. A custom-designed\nneural network is trained to estimate the optimal value and feasibility of the\nsecond-stage problem. The network can be incorporated into CCG, leading to more\ncomputationally tractable subproblems in each of its iterations. The resulting\nalgorithm enjoys approximation guarantees which depend on the neural network's\nprediction error. In our experiments, Neur2RO produces high-quality solutions\nquickly, outperforming state-of-the-art methods on two-stage knapsack, capital\nbudgeting, and facility location problems. Compared to existing methods, which\noften run for hours, Neur2RO finds better solutions in a few seconds or\nminutes. Our code is available at https://github.com/khalil-research/Neur2RO.\n", "date": "2023-10-06", "categories": []} {"id": "2403.15180", "title": "Self-Improvement for Neural Combinatorial Optimization: Sample without\n Replacement, but Improvement", "abstract": " Current methods for end-to-end constructive neural combinatorial optimization\nusually train a policy using behavior cloning from expert solutions or policy\ngradient methods from reinforcement learning. While behavior cloning is\nstraightforward, it requires expensive expert solutions, and policy gradient\nmethods are often computationally demanding and complex to fine-tune. In this\nwork, we bridge the two and simplify the training process by sampling multiple\nsolutions for random instances using the current model in each epoch and then\nselecting the best solution as an expert trajectory for supervised imitation\nlearning. To achieve progressively improving solutions with minimal sampling,\nwe introduce a method that combines round-wise Stochastic Beam Search with an\nupdate strategy derived from a provable policy improvement. This strategy\nrefines the policy between rounds by utilizing the advantage of the sampled\nsequences with almost no computational overhead. We evaluate our approach on\nthe Traveling Salesman Problem and the Capacitated Vehicle Routing Problem. The\nmodels trained with our method achieve comparable performance and\ngeneralization to those trained with expert data. Additionally, we apply our\nmethod to the Job Shop Scheduling Problem using a transformer-based\narchitecture and outperform existing state-of-the-art methods by a wide margin.\n", "date": "2024-03-22", "categories": []} {"id": "2407.17206", "title": "Take a Step and Reconsider: Sequence Decoding for Self-Improved Neural\n Combinatorial Optimization", "abstract": " The constructive approach within Neural Combinatorial Optimization (NCO)\ntreats a combinatorial optimization problem as a finite Markov decision\nprocess, where solutions are built incrementally through a sequence of\ndecisions guided by a neural policy network. To train the policy, recent\nresearch is shifting toward a 'self-improved' learning methodology that\naddresses the limitations of reinforcement learning and supervised approaches.\nHere, the policy is iteratively trained in a supervised manner, with solutions\nderived from the current policy serving as pseudo-labels. The way these\nsolutions are obtained from the policy determines the quality of the\npseudo-labels. In this paper, we present a simple and problem-independent\nsequence decoding method for self-improved learning based on sampling sequences\nwithout replacement. We incrementally follow the best solution found and repeat\nthe sampling process from intermediate partial solutions. By modifying the\npolicy to ignore previously sampled sequences, we force it to consider only\nunseen alternatives, thereby increasing solution diversity. Experimental\nresults for the Traveling Salesman and Capacitated Vehicle Routing Problem\ndemonstrate its strong performance. Furthermore, our method outperforms\nprevious NCO approaches on the Job Shop Scheduling Problem.\n", "date": "2024-07-24", "categories": []} {"id": "2309.16965", "title": "Controlling Continuous Relaxation for Combinatorial Optimization", "abstract": " Unsupervised learning (UL)-based solvers for combinatorial optimization (CO)\ntrain a neural network that generates a soft solution by directly optimizing\nthe CO objective using a continuous relaxation strategy. These solvers offer\nseveral advantages over traditional methods and other learning-based methods,\nparticularly for large-scale CO problems. However, UL-based solvers face two\npractical issues: (I) an optimization issue, where UL-based solvers are easily\ntrapped at local optima, and (II) a rounding issue, where UL-based solvers\nrequire artificial post-learning rounding from the continuous space back to the\noriginal discrete space, undermining the robustness of the results. This study\nproposes a Continuous Relaxation Annealing (CRA) strategy, an effective\nrounding-free learning method for UL-based solvers. CRA introduces a penalty\nterm that dynamically shifts from prioritizing continuous solutions,\neffectively smoothing the non-convexity of the objective function, to enforcing\ndiscreteness, eliminating artificial rounding. Experimental results demonstrate\nthat CRA significantly enhances the performance of UL-based solvers,\noutperforming existing UL-based solvers and greedy algorithms in complex CO\nproblems. Additionally, CRA effectively eliminates artificial rounding and\naccelerates the learning process.\n", "date": "2023-09-29", "categories": []} {"id": "2401.11849", "title": "Self-Labeling the Job Shop Scheduling Problem", "abstract": " This work proposes a self-supervised training strategy designed for\ncombinatorial problems. An obstacle in applying supervised paradigms to such\nproblems is the need for costly target solutions often produced with exact\nsolvers. Inspired by semi- and self-supervised learning, we show that\ngenerative models can be trained by sampling multiple solutions and using the\nbest one according to the problem objective as a pseudo-label. In this way, we\niteratively improve the model generation capability by relying only on its\nself-supervision, eliminating the need for optimality information. We validate\nthis Self-Labeling Improvement Method (SLIM) on the Job Shop Scheduling (JSP),\na complex combinatorial problem that is receiving much attention from the\nneural combinatorial community. We propose a generative model based on the\nwell-known Pointer Network and train it with SLIM. Experiments on popular\nbenchmarks demonstrate the potential of this approach as the resulting models\noutperform constructive heuristics and state-of-the-art learning proposals for\nthe JSP. Lastly, we prove the robustness of SLIM to various parameters and its\ngenerality by applying it to the Traveling Salesman Problem.\n", "date": "2024-01-22", "categories": []} {"id": "2410.22806", "title": "MILP-StuDio: MILP Instance Generation via Block Structure Decomposition", "abstract": " Mixed-integer linear programming (MILP) is one of the most popular\nmathematical formulations with numerous applications. In practice, improving\nthe performance of MILP solvers often requires a large amount of high-quality\ndata, which can be challenging to collect. Researchers thus turn to generation\ntechniques to generate additional MILP instances. However, existing approaches\ndo not take into account specific block structures -- which are closely related\nto the problem formulations -- in the constraint coefficient matrices (CCMs) of\nMILPs. Consequently, they are prone to generate computationally trivial or\ninfeasible instances due to the disruptions of block structures and thus\nproblem formulations. To address this challenge, we propose a novel MILP\ngeneration framework, called Block Structure Decomposition (MILP-StuDio), to\ngenerate high-quality instances by preserving the block structures.\nSpecifically, MILP-StuDio begins by identifying the blocks in CCMs and\ndecomposing the instances into block units, which serve as the building blocks\nof MILP instances. We then design three operators to construct new instances by\nremoving, substituting, and appending block units in the original instances,\nenabling us to generate instances with flexible sizes. An appealing feature of\nMILP-StuDio is its strong ability to preserve the feasibility and computational\nhardness of the generated instances. Experiments on the commonly-used\nbenchmarks demonstrate that using instances generated by MILP-StuDio is able to\nsignificantly reduce over 10% of the solving time for learning-based solvers.\n", "date": "2024-10-30", "categories": []} {"id": "2410.23934", "title": "Towards Fast Algorithms for the Preference Consistency Problem Based on\n Hierarchical Models", "abstract": " In this paper, we construct and compare algorithmic approaches to solve the\nPreference Consistency Problem for preference statements based on hierarchical\nmodels. Instances of this problem contain a set of preference statements that\nare direct comparisons (strict and non-strict) between some alternatives, and a\nset of evaluation functions by which all alternatives can be rated. An instance\nis consistent based on hierarchical preference models, if there exists an\nhierarchical model on the evaluation functions that induces an order relation\non the alternatives by which all relations given by the preference statements\nare satisfied. Deciding if an instance is consistent is known to be NP-complete\nfor hierarchical models. We develop three approaches to solve this decision\nproblem. The first involves a Mixed Integer Linear Programming (MILP)\nformulation, the other two are recursive algorithms that are based on\nproperties of the problem by which the search space can be pruned. Our\nexperiments on synthetic data show that the recursive algorithms are faster\nthan solving the MILP formulation and that the ratio between the running times\nincreases extremely quickly.\n", "date": "2024-10-31", "categories": []} {"id": "2410.24052", "title": "Attention is All You Need to Optimize Wind Farm Operations and\n Maintenance", "abstract": " Operations and maintenance (O&M) is a fundamental problem in wind energy\nsystems with far reaching implications for reliability and profitability.\nOptimizing O&M is a multi-faceted decision optimization problem that requires a\ncareful balancing act across turbine level failure risks, operational revenues,\nand maintenance crew logistics. The resulting O&M problems are typically solved\nusing large-scale mixed integer programming (MIP) models, which yield\ncomputationally challenging problems that require either long-solution times,\nor heuristics to reach a solution. To address this problem, we introduce a\nnovel decision-making framework for wind farm O&M that builds on a multi-head\nattention (MHA) models, an emerging artificial intelligence methods that are\nspecifically designed to learn in rich and complex problem settings. The\ndevelopment of proposed MHA framework incorporates a number of modeling\ninnovations that allows explicit embedding of MIP models within an MHA\nstructure. The proposed MHA model (i) significantly reduces the solution time\nfrom hours to seconds, (ii) guarantees feasibility of the proposed solutions\nconsidering complex constraints that are omnipresent in wind farm O&M, (iii)\nresults in significant solution quality compared to the conventional MIP\nformulations, and (iv) exhibits significant transfer learning capability across\ndifferent problem settings.\n", "date": "2024-10-31", "categories": []} {"id": "2406.06954", "title": "Distributional MIPLIB: a Multi-Domain Library for Advancing ML-Guided\n MILP Methods", "abstract": " Mixed Integer Linear Programming (MILP) is a fundamental tool for modeling\ncombinatorial optimization problems. Recently, a growing body of research has\nused machine learning to accelerate MILP solving. Despite the increasing\npopularity of this approach, there is a lack of a common repository that\nprovides distributions of similar MILP instances across different domains, at\ndifferent hardness levels, with standardized test sets. In this paper, we\nintroduce Distributional MIPLIB, a multi-domain library of problem\ndistributions for advancing ML-guided MILP methods. We curate MILP\ndistributions from existing work in this area as well as real-world problems\nthat have not been used, and classify them into different hardness levels. It\nwill facilitate research in this area by enabling comprehensive evaluation on\ndiverse and realistic domains. We empirically illustrate the benefits of using\nDistributional MIPLIB as a research vehicle in two ways. We evaluate the\nperformance of ML-guided variable branching on previously unused distributions\nto identify potential areas for improvement. Moreover, we propose to learn\nbranching policies from a mix of distributions, demonstrating that mixed\ndistributions achieve better performance compared to homogeneous distributions\nwhen there is limited data and generalize well to larger instances. The dataset\nis publicly available at\nhttps://sites.google.com/usc.edu/distributional-miplib/home.\n", "date": "2024-06-11", "categories": []} {"id": "2410.21066", "title": "Learning to Handle Complex Constraints for Vehicle Routing Problems", "abstract": " Vehicle Routing Problems (VRPs) can model many real-world scenarios and often\ninvolve complex constraints. While recent neural methods excel in constructing\nsolutions based on feasibility masking, they struggle with handling complex\nconstraints, especially when obtaining the masking itself is NP-hard. In this\npaper, we propose a novel Proactive Infeasibility Prevention (PIP) framework to\nadvance the capabilities of neural methods towards more complex VRPs. Our PIP\nintegrates the Lagrangian multiplier as a basis to enhance constraint awareness\nand introduces preventative infeasibility masking to proactively steer the\nsolution construction process. Moreover, we present PIP-D, which employs an\nauxiliary decoder and two adaptive strategies to learn and predict these\ntailored masks, potentially enhancing performance while significantly reducing\ncomputational costs during training. To verify our PIP designs, we conduct\nextensive experiments on the highly challenging Traveling Salesman Problem with\nTime Window (TSPTW), and TSP with Draft Limit (TSPDL) variants under different\nconstraint hardness levels. Notably, our PIP is generic to boost many neural\nmethods, and exhibits both a significant reduction in infeasible rate and a\nsubstantial improvement in solution quality.\n", "date": "2024-10-28", "categories": []} {"id": "2311.02629", "title": "Pointer Networks with Q-Learning for Combinatorial Optimization", "abstract": " We introduce the Pointer Q-Network (PQN), a hybrid neural architecture that\nintegrates model-free Q-value policy approximation with Pointer Networks\n(Ptr-Nets) to enhance the optimality of attention-based sequence generation,\nfocusing on long-term outcomes. This integration proves particularly effective\nin solving combinatorial optimization (CO) tasks, especially the Travelling\nSalesman Problem (TSP), which is the focus of our study. We address this\nchallenge by defining a Markov Decision Process (MDP) compatible with PQN,\nwhich involves iterative graph embedding, encoding and decoding by an\nLSTM-based recurrent neural network. This process generates a context vector\nand computes raw attention scores, which are dynamically adjusted by Q-values\ncalculated for all available state-action pairs before applying softmax. The\nresulting attention vector is utilized as an action distribution, with actions\nselected hinged to exploration-exploitation dynamic adaptibility of PQN. Our\nempirical results demonstrate the efficacy of this approach, also testing the\nmodel in unstable environments.\n", "date": "2023-11-05", "categories": []} {"id": "2312.06314", "title": "DMS*: Minimizing Makespan for Multi-Agent Combinatorial Path Finding", "abstract": " Multi-Agent Combinatorial Path Finding (MCPF) seeks collision-free paths for\nmultiple agents from their initial to goal locations, while visiting a set of\nintermediate target locations in the middle of the paths. MCPF is challenging\nas it involves both planning collision-free paths for multiple agents and\ntarget sequencing, i.e., solving traveling salesman problems to assign targets\nto and find the visiting order for the agents. Recent work develops methods to\naddress MCPF while minimizing the sum of individual arrival times at goals.\nSuch a problem formulation may result in paths with different arrival times and\nlead to a long makespan, the maximum arrival time, among the agents. This paper\nproposes a min-max variant of MCPF, denoted as MCPF-max, that minimizes the\nmakespan of the agents. While the existing methods (such as MS*) for MCPF can\nbe adapted to solve MCPF-max, we further develop two new techniques based on\nMS* to defer the expensive target sequencing during planning to expedite the\noverall computation. We analyze the properties of the resulting algorithm\nDeferred MS* (DMS*), and test DMS* with up to 20 agents and 80 targets. We\ndemonstrate the use of DMS* on differential-drive robots.\n", "date": "2023-12-11", "categories": []} {"id": "2410.18786", "title": "Applying Neural Monte Carlo Tree Search to Unsignalized\n Multi-intersection Scheduling for Autonomous Vehicles", "abstract": " Dynamic scheduling of access to shared resources by autonomous systems is a\nchallenging problem, characterized as being NP-hard. The complexity of this\ntask leads to a combinatorial explosion of possibilities in highly dynamic\nsystems where arriving requests must be continuously scheduled subject to\nstrong safety and time constraints. An example of such a system is an\nunsignalized intersection, where automated vehicles' access to potential\nconflict zones must be dynamically scheduled. In this paper, we apply Neural\nMonte Carlo Tree Search (NMCTS) to the challenging task of scheduling platoons\nof vehicles crossing unsignalized intersections. Crucially, we introduce a\ntransformation model that maps successive sequences of potentially conflicting\nroad-space reservation requests from platoons of vehicles into a series of\nboard-game-like problems and use NMCTS to search for solutions representing\noptimal road-space allocation schedules in the context of past allocations. To\noptimize search, we incorporate a prioritized re-sampling method with parallel\nNMCTS (PNMCTS) to improve the quality of training data. To optimize training, a\ncurriculum learning strategy is used to train the agent to schedule\nprogressively more complex boards culminating in overlapping boards that\nrepresent busy intersections. In a busy single four-way unsignalized\nintersection simulation, PNMCTS solved 95\\% of unseen scenarios, reducing\ncrossing time by 43\\% in light and 52\\% in heavy traffic versus first-in,\nfirst-out control. In a 3x3 multi-intersection network, the proposed method\nmaintained free-flow in light traffic when all intersections are under control\nof PNMCTS and outperformed state-of-the-art RL-based traffic-light controllers\nin average travel time by 74.5\\% and total throughput by 16\\% in heavy traffic.\n", "date": "2024-10-24", "categories": []} {"id": "2410.17415", "title": "End-to-End Optimization and Learning of Fair Court Schedules", "abstract": " Criminal courts across the United States handle millions of cases every year,\nand the scheduling of those cases must accommodate a diverse set of\nconstraints, including the preferences and availability of courts, prosecutors,\nand defense teams. When criminal court schedules are formed, defendants'\nscheduling preferences often take the least priority, although defendants may\nface significant consequences (including arrest or detention) for missed court\ndates. Additionally, studies indicate that defendants' nonappearances impose\ncosts on the courts and other system stakeholders. To address these issues,\ncourts and commentators have begun to recognize that pretrial outcomes for\ndefendants and for the system would be improved with greater attention to court\nprocesses, including \\emph{court scheduling practices}. There is thus a need\nfor fair criminal court pretrial scheduling systems that account for\ndefendants' preferences and availability, but the collection of such data poses\nlogistical challenges. Furthermore, optimizing schedules fairly across various\nparties' preferences is a complex optimization problem, even when such data is\navailable. In an effort to construct such a fair scheduling system under data\nuncertainty, this paper proposes a joint optimization and learning framework\nthat combines machine learning models trained end-to-end with efficient\nmatching algorithms. This framework aims to produce court scheduling schedules\nthat optimize a principled measure of fairness, balancing the availability and\npreferences of all parties.\n", "date": "2024-10-22", "categories": []} {"id": "2410.17943", "title": "Optimizing Travel Itineraries with AI Algorithms in a Microservices\n Architecture: Balancing Cost, Time, Preferences, and Sustainability", "abstract": " The objective of this research is how an implementation of AI algorithms in\nthe microservices architecture enhances travel itineraries by cost, time, user\npreferences, and environmental sustainability. It uses machine learning models\nfor both cost forecasting and personalization, genetic algorithm for\noptimization of the itinerary, and heuristics for sustainability checking.\nPrimary evaluated parameters consist of latency, ability to satisfy user\npreferences, cost and environmental concern. The experimental results\ndemonstrate an average of 4.5 seconds of response time on 1000 concurrent users\nand 92% of user preferences accuracy. The cost efficiency is proved, with 95%\nof provided trips being within the limits of the budget declared by the user.\nThe system also implements some measures to alleviate negative externalities\nrelated to travel and 60% of offered travel plans had green options\nincorporated, resulting in the average 15% lower carbon emissions than the\ntraditional travel plans offered. The genetic algorithm with time complexity\nO(g.p.f) provides the optimal solution in 100 generations. Every iteration\nimproves the quality of the solution by 5%, thus enabling its effective use in\noptimization problems where time is measured in seconds. Finally, the system is\ndesigned to be fault-tolerant with functional 99.9% availability which allows\nthe provision of services even when requirements are exceeded. Travel\noptimization platform is turned dynamic and efficient by this microservices\nbased architecture which provides enhanced scaling, allows asynchronous\ncommunication and real time changes. Because of the incorporation of Ai, cost\ncontrol and eco-friendliness approaches, the system addresses the different\nuser needs in the present days travel business.\n", "date": "2024-10-23", "categories": []} {"id": "2410.17945", "title": "Theoretically Grounded Pruning of Large Ground Sets for Constrained,\n Discrete Optimization", "abstract": " Modern instances of combinatorial optimization problems often exhibit\nbillion-scale ground sets, which have many uninformative or redundant elements.\nIn this work, we develop light-weight pruning algorithms to quickly discard\nelements that are unlikely to be part of an optimal solution. Under mild\nassumptions on the instance, we prove theoretical guarantees on the fraction of\nthe optimal value retained and the size of the resulting pruned ground set.\nThrough extensive experiments on real-world datasets for various applications,\nwe demonstrate that our algorithm, QuickPrune, efficiently prunes over 90% of\nthe ground set and outperforms state-of-the-art classical and machine learning\nheuristics for pruning.\n", "date": "2024-10-23", "categories": []} {"id": "2410.13226", "title": "Research on Travel Route Planing Problems Based on Greedy Algorithm", "abstract": " The route planning problem based on the greedy algorithm represents a method\nof identifying the optimal or near-optimal route between a given start point\nand end point. In this paper, the PCA method is employed initially to downscale\nthe city evaluation indexes, extract the key principal components, and then\ndownscale the data using the KMO and TOPSIS algorithms, all of which are based\non the MindSpore framework. Secondly, for the dataset that does not pass the\nKMO test, the entropy weight method and TOPSIS method will be employed for\ncomprehensive evaluation. Finally, a route planning algorithm is proposed and\noptimised based on the greedy algorithm, which provides personalised route\ncustomisation according to the different needs of tourists. In addition, the\nlocal travelling efficiency, the time required to visit tourist attractions and\nthe necessary daily breaks are considered in order to reduce the cost and avoid\nfalling into the locally optimal solution.\n", "date": "2024-10-17", "categories": []} {"id": "2410.17111", "title": "Permutation Picture of Graph Combinatorial Optimization Problems", "abstract": " This paper proposes a framework that formulates a wide range of graph\ncombinatorial optimization problems using permutation-based representations.\nThese problems include the travelling salesman problem, maximum independent\nset, maximum cut, and various other related problems. This work potentially\nopens up new avenues for algorithm design in neural combinatorial optimization,\nbridging the gap between discrete and continuous optimization techniques.\n", "date": "2024-10-22", "categories": []} {"id": "2410.17160", "title": "Layered LA-MAPF: a decomposition of large agent MAPF instance to\n accelerate solving without compromising solvability", "abstract": " Multi-Agent Path Finding (MAPF) has been widely studied in recent years.\nHowever, most existing MAPF algorithms assume that an agent occupies only a\nsingle grid in a grid-based map. This assumption limits their applicability in\nmany real-world domains where agents have geometric shapes, rather than being\npoint-like. Such agents, which can occupy multiple cells simultaneously, are\nreferred to as ``large'' agents. When considering the shape and size of agents\nin MAPF, the computational complexity increases significantly as the number of\nagents grows, primarily due to the increased overhead in conflict detection\nbetween geometric agents. In this paper, we propose two types of subproblems\nfor the LA-MAPF (Large-Agent MAPF) problem: \\textbf{cluster} (which has no\nconstraints on the order of solution) and \\textbf{level} (which imposes\nconstraints on the solution order). We introduce \\textbf{Layered LA-MAPF}, a\nmethod that decomposes a MAPF instance involving geometric agents into\nclusters, and then further decomposes each cluster into levels. This approach\naims to reduce time complexity when solving LA-MAPF problems. Our results\ndemonstrate the performance of our method as the number of agents increases\nacross various maps, and how it accelerates LA-MAPF methods, such as LA-CBS and\nLA-LaCAM. Experiments show that our LA-MAPF method with instance decomposition\n\\textbf{halves the time cost (reducing from an average of 40s to 20s) and\ntriples the success rate (from an average of 0.27 to 0.80)} in finding a\nsolution within 60 seconds. To facilitate further research, we have made the\nsource code for Layered LA-MAPF publicly available at\n\\url{https://github.com/JoeYao-bit/LayeredMAPF/algorithm/LA-MAPF}.\n", "date": "2024-10-22", "categories": []} {"id": "2410.14947", "title": "Optimally Solving Colored Generalized Sliding-Tile Puzzles: Complexity\n and Bounds", "abstract": " The Generalized Sliding-Tile Puzzle (GSTP), allowing many square tiles on a\nboard to move in parallel while enforcing natural geometric collision\nconstraints on the movement of neighboring tiles, provide a high-fidelity\nmathematical model for many high-utility existing and future multi-robot\napplications, e.g., at mobile robot-based warehouses or autonomous garages.\nMotivated by practical relevance, this work examines a further generalization\nof GSTP called the Colored Generalized Sliding-Tile Puzzle (CGSP), where tiles\ncan now assume varying degrees of distinguishability, a common occurrence in\nthe aforementioned applications. Our study establishes the computational\ncomplexity of CGSP and its key sub-problems under a broad spectrum of possible\nconditions and characterizes solution makespan lower and upper bounds that\ndiffer by at most a logarithmic factor. These results are further extended to\nhigher-dimensional versions of the puzzle game.\n", "date": "2024-10-19", "categories": []} {"id": "2410.15420", "title": "Where to Build Food Banks and Pantries: A Two-Level Machine Learning\n Approach", "abstract": " Over 44 million Americans currently suffer from food insecurity, of whom 13\nmillion are children. Across the United States, thousands of food banks and\npantries serve as vital sources of food and other forms of aid for food\ninsecure families. By optimizing food bank and pantry locations, food would\nbecome more accessible to families who desperately require it. In this work, we\nintroduce a novel two-level optimization framework, which utilizes the\nK-Medoids clustering algorithm in conjunction with the Open-Source Routing\nMachine engine, to optimize food bank and pantry locations based on real road\ndistances to houses and house blocks. Our proposed framework also has the\nadaptability to factor in considerations such as median household income using\na pseudo-weighted K-Medoids algorithm. Testing conducted with California and\nIndiana household data, as well as comparisons with real food bank and pantry\nlocations showed that interestingly, our proposed framework yields food pantry\nlocations superior to those of real existing ones and saves significant\ndistance for households, while there is a marginal penalty on the first level\nfood bank to food pantry distance. Overall, we believe that the second-level\nbenefits of this framework far outweigh any drawbacks and yield a net benefit\nresult.\n", "date": "2024-10-20", "categories": []} {"id": "2410.15601", "title": "All You Need is an Improving Column: Enhancing Column Generation for\n Parallel Machine Scheduling via Transformers", "abstract": " We present a neural network-enhanced column generation (CG) approach for a\nparallel machine scheduling problem. The proposed approach utilizes an\nencoder-decoder attention model, namely the transformer and pointer\narchitectures, to develop job sequences with negative reduced cost and thus\ngenerate columns to add to the master problem. By training the neural network\noffline and using it in inference mode to predict negative reduced costs\ncolumns, we achieve significant computational time savings compared to dynamic\nprogramming (DP). Since the exact DP procedure is used to verify that no\nfurther columns with negative reduced cost can be identified at termination,\nthe optimality guarantee of the original CG procedure is preserved. For small\nto medium-sized instances, our approach achieves an average 45% reduction in\ncomputation time compared to solving the subproblems with DP. Furthermore, the\nmodel generalizes not only to unseen, larger problem instances from the same\nprobability distribution but also to instances from different probability\ndistributions than those presented at training time. For large-sized instances,\nthe proposed approach achieves an 80% improvement in the objective value in\nunder 500 seconds, demonstrating both its scalability and efficiency.\n", "date": "2024-10-21", "categories": []} {"id": "2310.14659", "title": "Predicting Accurate Lagrangian Multipliers for Mixed Integer Linear\n Programs", "abstract": " Lagrangian relaxation stands among the most efficient approaches for solving\na Mixed Integer Linear Programs (MILP) with difficult constraints. Given any\nduals for these constraints, called Lagrangian Multipliers (LMs), it returns a\nbound on the optimal value of the MILP, and Lagrangian methods seek the LMs\ngiving the best such bound. But these methods generally rely on iterative\nalgorithms resembling gradient descent to maximize the concave piecewise linear\ndual function: the computational burden grows quickly with the number of\nrelaxed constraints. We introduce a deep learning approach that bypasses the\ndescent, effectively amortizing the local, per instance, optimization. A\nprobabilistic encoder based on a graph convolutional network computes\nhigh-dimensional representations of relaxed constraints in MILP instances. A\ndecoder then turns these representations into LMs. We train the encoder and\ndecoder jointly by directly optimizing the bound obtained from the predicted\nmultipliers. Numerical experiments show that our approach closes up to 85~\\% of\nthe gap between the continuous relaxation and the best Lagrangian bound, and\nprovides a high quality warm-start for descent based Lagrangian methods.\n", "date": "2023-10-23", "categories": []} {"id": "2410.12913", "title": "Fair Clustering for Data Summarization: Improved Approximation\n Algorithms and Complexity Insights", "abstract": " Data summarization tasks are often modeled as $k$-clustering problems, where\nthe goal is to choose $k$ data points, called cluster centers, that best\nrepresent the dataset by minimizing a clustering objective. A popular objective\nis to minimize the maximum distance between any data point and its nearest\ncenter, which is formalized as the $k$-center problem. While in some\napplications all data points can be chosen as centers, in the general setting,\ncenters must be chosen from a predefined subset of points, referred as\nfacilities or suppliers; this is known as the $k$-supplier problem. In this\nwork, we focus on fair data summarization modeled as the fair $k$-supplier\nproblem, where data consists of several groups, and a minimum number of centers\nmust be selected from each group while minimizing the $k$-supplier objective.\nThe groups can be disjoint or overlapping, leading to two distinct problem\nvariants each with different computational complexity.\n We present $3$-approximation algorithms for both variants, improving the\npreviously known factor of $5$. For disjoint groups, our algorithm runs in\npolynomial time, while for overlapping groups, we present a fixed-parameter\ntractable algorithm, where the exponential runtime depends only on the number\nof groups and centers. We show that these approximation factors match the\ntheoretical lower bounds, assuming standard complexity theory conjectures.\nFinally, using an open-source implementation, we demonstrate the scalability of\nour algorithms on large synthetic datasets and assess the price of fairness on\nreal-world data, comparing solution quality with and without fairness\nconstraints.\n", "date": "2024-10-16", "categories": []} {"id": "2308.00560", "title": "Reinforcement Learning-based Non-Autoregressive Solver for Traveling\n Salesman Problems", "abstract": " The Traveling Salesman Problem (TSP) is a well-known combinatorial\noptimization problem with broad real-world applications. Recently, neural\nnetworks have gained popularity in this research area because as shown in the\nliterature, they provide strong heuristic solutions to TSPs. Compared to\nautoregressive neural approaches, non-autoregressive (NAR) networks exploit the\ninference parallelism to elevate inference speed but suffer from comparatively\nlow solution quality. In this paper, we propose a novel NAR model named\nNAR4TSP, which incorporates a specially designed architecture and an enhanced\nreinforcement learning strategy. To the best of our knowledge, NAR4TSP is the\nfirst TSP solver that successfully combines RL and NAR networks. The key lies\nin the incorporation of NAR network output decoding into the training process.\nNAR4TSP efficiently represents TSP encoded information as rewards and\nseamlessly integrates it into reinforcement learning strategies, while\nmaintaining consistent TSP sequence constraints during both training and\ntesting phases. Experimental results on both synthetic and real-world TSPs\ndemonstrate that NAR4TSP outperforms five state-of-the-art models in terms of\nsolution quality, inference speed, and generalization to unseen scenarios.\n", "date": "2023-08-01", "categories": []} {"id": "2410.12031", "title": "A Learning Search Algorithm for the Restricted Longest Common\n Subsequence Problem", "abstract": " This paper addresses the Restricted Longest Common Subsequence (RLCS)\nproblem, an extension of the well-known Longest Common Subsequence (LCS)\nproblem. This problem has significant applications in bioinformatics,\nparticularly for identifying similarities and discovering mutual patterns and\nimportant motifs among DNA, RNA, and protein sequences. Building on recent\nadvancements in solving this problem through a general search framework, this\npaper introduces two novel heuristic approaches designed to enhance the search\nprocess by steering it towards promising regions in the search space. The first\nheuristic employs a probabilistic model to evaluate partial solutions during\nthe search process. The second heuristic is based on a neural network model\ntrained offline using a genetic algorithm. A key aspect of this approach is\nextracting problem-specific features of partial solutions and the complete\nproblem instance. An effective hybrid method, referred to as the learning beam\nsearch, is developed by combining the trained neural network model with a beam\nsearch framework. An important contribution of this paper is found in the\ngeneration of real-world instances where scientific abstracts serve as input\nstrings, and a set of frequently occurring academic words from the literature\nare used as restricted patterns. Comprehensive experimental evaluations\ndemonstrate the effectiveness of the proposed approaches in solving the RLCS\nproblem. Finally, an empirical explainability analysis is applied to the\nobtained results. In this way, key feature combinations and their respective\ncontributions to the success or failure of the algorithms across different\nproblem types are identified.\n", "date": "2024-10-15", "categories": []} {"id": "2210.11152", "title": "Machine Learning for K-adaptability in Two-stage Robust Optimization", "abstract": " Two-stage robust optimization problems constitute one of the hardest\noptimization problem classes. One of the solution approaches to this class of\nproblems is K-adaptability. This approach simultaneously seeks the best\npartitioning of the uncertainty set of scenarios into K subsets, and optimizes\ndecisions corresponding to each of these subsets. In general case, it is solved\nusing the K-adaptability branch-and-bound algorithm, which requires exploration\nof exponentially-growing solution trees. To accelerate finding high-quality\nsolutions in such trees, we propose a machine learning-based node selection\nstrategy. In particular, we construct a feature engineering scheme based on\ngeneral two-stage robust optimization insights that allows us to train our\nmachine learning tool on a database of resolved B&B trees, and to apply it\nas-is to problems of different sizes and/or types. We experimentally show that\nusing our learned node selection strategy outperforms a vanilla, random node\nselection strategy when tested on problems of the same type as the training\nproblems, also in case the K-value or the problem size differs from the\ntraining ones.\n", "date": "2022-10-20", "categories": []} {"id": "2211.00759", "title": "Online Control of Adaptive Large Neighborhood Search using Deep\n Reinforcement Learning", "abstract": " The Adaptive Large Neighborhood Search (ALNS) algorithm has shown\nconsiderable success in solving combinatorial optimization problems (COPs).\nNonetheless, the performance of ALNS relies on the proper configuration of its\nselection and acceptance parameters, which is known to be a complex and\nresource-intensive task. To address this, we introduce a Deep Reinforcement\nLearning (DRL) based approach called DR-ALNS that selects operators, adjusts\nparameters, and controls the acceptance criterion throughout the search. The\nproposed method aims to learn, based on the state of the search, to configure\nALNS for the next iteration to yield more effective solutions for the given\noptimization problem. We evaluate the proposed method on an orienteering\nproblem with stochastic weights and time windows, as presented in an IJCAI\ncompetition. The results show that our approach outperforms vanilla ALNS, ALNS\ntuned with Bayesian optimization, and two state-of-the-art DRL approaches that\nwere the winning methods of the competition, achieving this with significantly\nfewer training observations. Furthermore, we demonstrate several good\nproperties of the proposed DR-ALNS method: it is easily adapted to solve\ndifferent routing problems, its learned policies perform consistently well\nacross various instance sizes, and these policies can be directly applied to\ndifferent problem variants.\n", "date": "2022-11-01", "categories": []} {"id": "2402.01145", "title": "ReEvo: Large Language Models as Hyper-Heuristics with Reflective\n Evolution", "abstract": " The omnipresence of NP-hard combinatorial optimization problems (COPs)\ncompels domain experts to engage in trial-and-error heuristic design. The\nlong-standing endeavor of design automation has gained new momentum with the\nrise of large language models (LLMs). This paper introduces Language\nHyper-Heuristics (LHHs), an emerging variant of Hyper-Heuristics that leverages\nLLMs for heuristic generation, featuring minimal manual intervention and\nopen-ended heuristic spaces. To empower LHHs, we present Reflective Evolution\n(ReEvo), a novel integration of evolutionary search for efficiently exploring\nthe heuristic space, and LLM reflections to provide verbal gradients within the\nspace. Across five heterogeneous algorithmic types, six different COPs, and\nboth white-box and black-box views of COPs, ReEvo yields state-of-the-art and\ncompetitive meta-heuristics, evolutionary algorithms, heuristics, and neural\nsolvers, while being more sample-efficient than prior LHHs.\n", "date": "2024-02-02", "categories": []} {"id": "2410.09173", "title": "Resource-Constrained Heuristic for Max-SAT", "abstract": " We propose a resource-constrained heuristic for instances of Max-SAT that\niteratively decomposes a larger problem into smaller subcomponents that can be\nsolved by optimized solvers and hardware. The unconstrained outer loop\nmaintains the state space of a given problem and selects a subset of the SAT\nvariables for optimization independent of previous calls. The\nresource-constrained inner loop maximizes the number of satisfiable clauses in\nthe \"sub-SAT\" problem. Our outer loop is agnostic to the mechanisms of the\ninner loop, allowing for the use of traditional solvers for the optimization\nstep. However, we can also transform the selected \"sub-SAT\" problem into a\nquadratic unconstrained binary optimization (QUBO) one and use specialized\nhardware for optimization. In contrast to existing solutions that convert a SAT\ninstance into a QUBO one before decomposition, we choose a subset of the SAT\nvariables before QUBO optimization. We analyze a set of variable selection\nmethods, including a novel graph-based method that exploits the structure of a\ngiven SAT instance. The number of QUBO variables needed to encode a (sub-)SAT\nproblem varies, so we additionally learn a model that predicts the size of\nsub-SAT problems that will fit a fixed-size QUBO solver. We empirically\ndemonstrate our results on a set of randomly generated Max-SAT instances as\nwell as real world examples from the Max-SAT evaluation benchmarks and\noutperform existing QUBO decomposer solutions.\n", "date": "2024-10-11", "categories": []} {"id": "2410.09316", "title": "Combinatorial optimization of the coefficient of determination", "abstract": " Robust correlation analysis is among the most critical challenges in\nstatistics. Herein, we develop an efficient algorithm for selecting the $k$-\nsubset of $n$ points in the plane with the highest coefficient of determination\n$\\left( R^2 \\right)$. Drawing from combinatorial geometry, we propose a method\ncalled the \\textit{quadratic sweep} that consists of two steps: (i)\nprojectively lifting the data points into $\\mathbb R^5$ and then (ii) iterating\nover each linearly separable $k$-subset. Its basis is that the optimal set of\noutliers is separable from its complement in $\\mathbb R^2$ by a conic section,\nwhich, in $\\mathbb R^5$, can be found by a topological sweep in $\\Theta \\left(\nn^5 \\log n \\right)$ time. Although key proofs of quadratic separability remain\nunderway, we develop strong mathematical intuitions for our conjectures, then\nexperimentally demonstrate our method's optimality over several million trials\nup to $n=30$ without error. Implementations in Julia and fully seeded,\nreproducible experiments are available at\nhttps://github.com/marc-harary/QuadraticSweep.\n", "date": "2024-10-12", "categories": []} {"id": "2410.08841", "title": "Public Transport Network Design for Equality of Accessibility via\n Message Passing Neural Networks and Reinforcement Learning", "abstract": " Designing Public Transport (PT) networks able to satisfy mobility needs of\npeople is essential to reduce the number of individual vehicles on the road,\nand thus pollution and congestion. Urban sustainability is thus tightly coupled\nto an efficient PT. Current approaches on Transport Network Design (TND)\ngenerally aim to optimize generalized cost, i.e., a unique number including\noperator and users' costs. Since we intend quality of PT as the capability of\nsatisfying mobility needs, we focus instead on PT accessibility, i.e., the ease\nof reaching surrounding points of interest via PT. PT accessibility is\ngenerally unequally distributed in urban regions: suburbs generally suffer from\npoor PT accessibility, which condemns residents therein to be dependent on\ntheir private cars. We thus tackle the problem of designing bus lines so as to\nminimize the inequality in the geographical distribution of accessibility. We\ncombine state-of-the-art Message Passing Neural Networks (MPNN) and\nReinforcement Learning. We show the efficacy of our method against\nmetaheuristics (classically used in TND) in a use case representing in\nsimplified terms the city of Montreal.\n", "date": "2024-10-11", "categories": []} {"id": "2310.00526", "title": "Are Graph Neural Networks Optimal Approximation Algorithms?", "abstract": " In this work we design graph neural network architectures that capture\noptimal approximation algorithms for a large class of combinatorial\noptimization problems, using powerful algorithmic tools from semidefinite\nprogramming (SDP). Concretely, we prove that polynomial-sized message-passing\nalgorithms can represent the most powerful polynomial time algorithms for Max\nConstraint Satisfaction Problems assuming the Unique Games Conjecture. We\nleverage this result to construct efficient graph neural network architectures,\nOptGNN, that obtain high-quality approximate solutions on landmark\ncombinatorial optimization problems such as Max-Cut, Min-Vertex-Cover, and\nMax-3-SAT. Our approach achieves strong empirical results across a wide range\nof real-world and synthetic datasets against solvers and neural baselines.\nFinally, we take advantage of OptGNN's ability to capture convex relaxations to\ndesign an algorithm for producing bounds on the optimal solution from the\nlearned embeddings of OptGNN.\n", "date": "2023-10-01", "categories": []} {"id": "2403.07917", "title": "A Neural-Evolutionary Algorithm for Autonomous Transit Network Design", "abstract": " Planning a public transit network is a challenging optimization problem, but\nessential in order to realize the benefits of autonomous buses. We propose a\nnovel algorithm for planning networks of routes for autonomous buses. We first\ntrain a graph neural net model as a policy for constructing route networks, and\nthen use the policy as one of several mutation operators in a evolutionary\nalgorithm. We evaluate this algorithm on a standard set of benchmarks for\ntransit network design, and find that it outperforms the learned policy alone\nby up to 20% and a plain evolutionary algorithm approach by up to 53% on\nrealistic benchmark instances.\n", "date": "2024-02-27", "categories": []} {"id": "2405.20042", "title": "CycleFormer : TSP Solver Based on Language Modeling", "abstract": " We propose a new transformer model for the Traveling Salesman Problem (TSP)\ncalled CycleFormer. We identified distinctive characteristics that need to be\nconsidered when applying a conventional transformer model to TSP and aimed to\nfully incorporate these elements into the TSP-specific transformer. Unlike the\ntoken sets in typical language models, which are limited and static, the token\n(node) set in TSP is unlimited and dynamic. To exploit this fact to the\nfullest, we equated the encoder output with the decoder linear layer and\ndirectly connected the context vector of the encoder to the decoder encoding.\nAdditionally, we added a positional encoding to the encoder tokens that\nreflects the two-dimensional nature of TSP, and devised a circular positional\nencoding for the decoder tokens that considers the cyclic properties of a tour.\nBy incorporating these ideas, CycleFormer outperforms state-of-the-art (SOTA)\ntransformer models for TSP from TSP-50 to TSP-500. Notably, on TSP-500, the\noptimality gap was reduced by approximately 2.8 times, from 3.09% to 1.10%,\ncompared to the existing SOTA. The code will be made available at\nhttps://github.com/Giventicket/CycleFormer.\n", "date": "2024-05-30", "categories": []} {"id": "2410.04968", "title": "Collaboration! Towards Robust Neural Methods for Routing Problems", "abstract": " Despite enjoying desirable efficiency and reduced reliance on domain\nexpertise, existing neural methods for vehicle routing problems (VRPs) suffer\nfrom severe robustness issues -- their performance significantly deteriorates\non clean instances with crafted perturbations. To enhance robustness, we\npropose an ensemble-based Collaborative Neural Framework (CNF) w.r.t. the\ndefense of neural VRP methods, which is crucial yet underexplored in the\nliterature. Given a neural VRP method, we adversarially train multiple models\nin a collaborative manner to synergistically promote robustness against\nattacks, while boosting standard generalization on clean instances. A neural\nrouter is designed to adeptly distribute training instances among models,\nenhancing overall load balancing and collaborative efficacy. Extensive\nexperiments verify the effectiveness and versatility of CNF in defending\nagainst various attacks across different neural VRP methods. Notably, our\napproach also achieves impressive out-of-distribution generalization on\nbenchmark instances.\n", "date": "2024-10-07", "categories": []} {"id": "2410.03205", "title": "A Tutorial on the Design, Experimentation and Application of\n Metaheuristic Algorithms to Real-World Optimization Problems", "abstract": " In the last few years, the formulation of real-world optimization problems\nand their efficient solution via metaheuristic algorithms has been a catalyst\nfor a myriad of research studies. In spite of decades of historical\nadvancements on the design and use of metaheuristics, large difficulties still\nremain in regards to the understandability, algorithmic design uprightness, and\nperformance verifiability of new technical achievements. A clear example stems\nfrom the scarce replicability of works dealing with metaheuristics used for\noptimization, which is often infeasible due to ambiguity and lack of detail in\nthe presentation of the methods to be reproduced. Additionally, in many cases,\nthere is a questionable statistical significance of their reported results.\nThis work aims at providing the audience with a proposal of good practices\nwhich should be embraced when conducting studies about metaheuristics methods\nused for optimization in order to provide scientific rigor, value and\ntransparency. To this end, we introduce a step by step methodology covering\nevery research phase that should be followed when addressing this scientific\nfield. Specifically, frequently overlooked yet crucial aspects and useful\nrecommendations will be discussed in regards to the formulation of the problem,\nsolution encoding, implementation of search operators, evaluation metrics,\ndesign of experiments, and considerations for real-world performance, among\nothers. Finally, we will outline important considerations, challenges, and\nresearch directions for the success of newly developed optimization\nmetaheuristics in their deployment and operation over real-world application\nenvironments.\n", "date": "2024-10-04", "categories": []} {"id": "2401.12478", "title": "Mini-batch Submodular Maximization", "abstract": " We present the first mini-batch algorithm for maximizing a non-negative\nmonotone decomposable submodular function, $F=\\sum_{i=1}^N f^i$, under a set of\nconstraints. We consider two sampling approaches: uniform and weighted. We\nfirst show that mini-batch with weighted sampling improves over the state of\nthe art sparsifier based approach both in theory and in practice.\n Surprisingly, our experimental results show that uniform sampling is superior\nto weighted sampling. However, it is impossible to explain this using\nworst-case analysis. Our main contribution is using smoothed analysis to\nprovide a theoretical foundation for our experimental results. We show that,\nunder very mild assumptions, uniform sampling is superior for both the\nmini-batch and the sparsifier approaches. We empirically verify that these\nassumptions hold for our datasets. Uniform sampling is simple to implement and\nhas complexity independent of $N$, making it the perfect candidate to tackle\nmassive real-world datasets.\n", "date": "2024-01-23", "categories": []} {"id": "2409.02135", "title": "Optimization by Parallel Quasi-Quantum Annealing with Gradient-Based\n Sampling", "abstract": " Learning-based methods have gained attention as general-purpose solvers due\nto their ability to automatically learn problem-specific heuristics, reducing\nthe need for manually crafted heuristics. However, these methods often face\nscalability challenges. To address these issues, the improved Sampling\nalgorithm for Combinatorial Optimization (iSCO), using discrete Langevin\ndynamics, has been proposed, demonstrating better performance than several\nlearning-based solvers. This study proposes a different approach that\nintegrates gradient-based update through continuous relaxation, combined with\nQuasi-Quantum Annealing (QQA). QQA smoothly transitions the objective function,\nstarting from a simple convex function, minimized at half-integral values, to\nthe original objective function, where the relaxed variables are minimized only\nin the discrete space. Furthermore, we incorporate parallel run communication\nleveraging GPUs to enhance exploration capabilities and accelerate convergence.\nNumerical experiments demonstrate that our method is a competitive\ngeneral-purpose solver, achieving performance comparable to iSCO and\nlearning-based solvers across various benchmark problems. Notably, our method\nexhibits superior speed-quality trade-offs for large-scale instances compared\nto iSCO, learning-based solvers, commercial solvers, and specialized\nalgorithms.\n", "date": "2024-09-02", "categories": []} {"id": "2410.01368", "title": "Theoretical Lower Bounds for the Oven Scheduling Problem", "abstract": " The Oven Scheduling Problem (OSP) is an NP-hard real-world parallel batch\nscheduling problem arising in the semiconductor industry. The objective of the\nproblem is to schedule a set of jobs on ovens while minimizing several factors,\nnamely total oven runtime, job tardiness, and setup costs. At the same time, it\nmust adhere to various constraints such as oven eligibility and availability,\njob release dates, setup times between batches, and oven capacity limitations.\nThe key to obtaining efficient schedules is to process compatible jobs\nsimultaneously in batches. In this paper, we develop theoretical,\nproblem-specific lower bounds for the OSP that can be computed very quickly. We\nthoroughly examine these lower bounds, evaluating their quality and exploring\ntheir integration into existing solution methods. Specifically, we investigate\ntheir contribution to exact methods and a metaheuristic local search approach\nusing simulated annealing. Moreover, these problem-specific lower bounds enable\nus to assess the solution quality for large instances for which exact methods\noften fail to provide tight lower bounds.\n", "date": "2024-10-02", "categories": []} {"id": "2410.01583", "title": "Iterated Local Search with Linkage Learning", "abstract": " In pseudo-Boolean optimization, a variable interaction graph represents\nvariables as vertices, and interactions between pairs of variables as edges. In\nblack-box optimization, the variable interaction graph may be at least\npartially discovered by using empirical linkage learning techniques. These\nmethods never report false variable interactions, but they are computationally\nexpensive. The recently proposed local search with linkage learning discovers\nthe partial variable interaction graph as a side-effect of iterated local\nsearch. However, information about the strength of the interactions is not\nlearned by the algorithm. We propose local search with linkage learning 2,\nwhich builds a weighted variable interaction graph that stores information\nabout the strength of the interaction between variables. The weighted variable\ninteraction graph can provide new insights about the optimization problem and\nbehavior of optimizers. Experiments with NK landscapes, knapsack problem, and\nfeature selection show that local search with linkage learning 2 is able to\nefficiently build weighted variable interaction graphs. In particular,\nexperiments with feature selection show that the weighted variable interaction\ngraphs can be used for visualizing the feature interactions in machine\nlearning. Additionally, new transformation operators that exploit the\ninteractions between variables can be designed. We illustrate this ability by\nproposing a new perturbation operator for iterated local search.\n", "date": "2024-10-02", "categories": []} {"id": "2409.04495", "title": "Learning to Solve Combinatorial Optimization under Positive Linear\n Constraints via Non-Autoregressive Neural Networks", "abstract": " Combinatorial optimization (CO) is the fundamental problem at the\nintersection of computer science, applied mathematics, etc. The inherent\nhardness in CO problems brings up challenge for solving CO exactly, making\ndeep-neural-network-based solvers a research frontier. In this paper, we design\na family of non-autoregressive neural networks to solve CO problems under\npositive linear constraints with the following merits. First, the positive\nlinear constraint covers a wide range of CO problems, indicating that our\napproach breaks the generality bottleneck of existing non-autoregressive\nnetworks. Second, compared to existing autoregressive neural network solvers,\nour non-autoregressive networks have the advantages of higher efficiency and\npreserving permutation invariance. Third, our offline unsupervised learning has\nlower demand on high-quality labels, getting rid of the demand of optimal\nlabels in supervised learning. Fourth, our online differentiable search method\nsignificantly improves the generalizability of our neural network solver to\nunseen problems. We validate the effectiveness of this framework in solving\nrepresentative CO problems including facility location, max-set covering, and\ntraveling salesman problem. Our non-autoregressive neural solvers are\ncompetitive to and can be even superior to state-of-the-art solvers such as\nSCIP and Gurobi, especially when both efficiency and efficacy are considered.\nCode is available at https://github.com/Thinklab-SJTU/NAR-CO-Solver\n", "date": "2024-09-06", "categories": []} {"id": "2309.00834", "title": "Approximating Fair $k$-Min-Sum-Radii in Euclidean Space", "abstract": " The $k$-center problem is a classical clustering problem in which one is\nasked to find a partitioning of a point set $P$ into $k$ clusters such that the\nmaximum radius of any cluster is minimized. It is well-studied. But what if we\nadd up the radii of the clusters instead of only considering the cluster with\nmaximum radius? This natural variant is called the $k$-min-sum-radii problem.\nIt has become the subject of more and more interest in recent years, inspiring\nthe development of approximation algorithms for the $k$-min-sum-radii problem\nin its plain version as well as in constrained settings.\n We study the problem for Euclidean spaces $\\mathbb{R}^d$ of arbitrary\ndimension but assume the number $k$ of clusters to be constant. In this case, a\nPTAS for the problem is known (see Bandyapadhyay, Lochet and Saurabh, SoCG,\n2023). Our aim is to extend the knowledge base for $k$-min-sum-radii to the\ndomain of fair clustering. We study several group fairness constraints, such as\nthe one introduced by Chierichetti et al. (NeurIPS, 2017). In this model, input\npoints have an additional attribute (e.g., colors such as red and blue), and\nclusters have to preserve the ratio between different attribute values (e.g.,\nhave the same fraction of red and blue points as the ground set). Different\nvariants of this general idea have been studied in the literature. To the best\nof our knowledge, no approximative results for the fair $k$-min-sum-radii\nproblem are known, despite the immense amount of work on the related fair\n$k$-center problem.\n We propose a PTAS for the fair $k$-min-sum-radii problem in Euclidean spaces\nof arbitrary dimension for the case of constant $k$. To the best of our\nknowledge, this is the first PTAS for the problem. It works for different\nnotions of group fairness.\n", "date": "2023-09-02", "categories": []} {"id": "2409.19668", "title": "Local Search for Integer Quadratic Programming", "abstract": " Integer Quadratic Programming (IQP) is an important problem in operations\nresearch. Local search is a powerful method for solving hard problems, but the\nresearch on local search algorithms for IQP solving is still on its early\nstage. This paper develops an efficient local search solver for solving general\nIQP, called LS-IQCQP. We propose four new local search operators for IQP that\ncan handle quadratic terms in the objective function, constraints or both.\nFurthermore, a two-mode local search algorithm is introduced, utilizing newly\ndesigned scoring functions to enhance the search process. Experiments are\nconducted on standard IQP benchmarks QPLIB and MINLPLIB, comparing LS-IQCQP\nwith several state-of-the-art IQP solvers. Experimental results demonstrate\nthat LS-IQCQP is competitive with the most powerful commercial solver Gurobi\nand outperforms other state-of-the-art solvers. Moreover, LS-IQCQP has\nestablished 6 new records for QPLIB and MINLPLIB open instances.\n", "date": "2024-09-29", "categories": []} {"id": "2409.18631", "title": "Quantum Algorithms for Drone Mission Planning", "abstract": " Mission planning often involves optimising the use of ISR (Intelligence,\nSurveillance and Reconnaissance) assets in order to achieve a set of mission\nobjectives within allowed parameters subject to constraints. The missions of\ninterest here, involve routing multiple UAVs visiting multiple targets,\nutilising sensors to capture data relating to each target. Finding such\nsolutions is often an NP-Hard problem and cannot be solved efficiently on\nclassical computers. Furthermore, during the mission new constraints and\nobjectives may arise, requiring a new solution to be computed within a short\ntime period. To achieve this we investigate near term quantum algorithms that\nhave the potential to offer speed-ups against current classical methods. We\ndemonstrate how a large family of these problems can be formulated as a Mixed\nInteger Linear Program (MILP) and then converted to a Quadratic Unconstrained\nBinary Optimisation (QUBO). The formulation provided is versatile and can be\nadapted for many different constraints with clear qubit scaling provided. We\ndiscuss the results of solving the QUBO formulation using commercial quantum\nannealers and compare the solutions to current edge classical solvers. We also\nanalyse the results from solving the QUBO using Quantum Approximate\nOptimisation Algorithms (QAOA) and discuss their results. Finally, we also\nprovide efficient methods to encode to the problem into the Variational Quantum\nEigensolver (VQE) formalism, where we have tailored the ansatz to the problem\nmaking efficient use of the qubits available.\n", "date": "2024-09-27", "categories": []} {"id": "2409.18692", "title": "MG-Net: Learn to Customize QAOA with Circuit Depth Awareness", "abstract": " Quantum Approximate Optimization Algorithm (QAOA) and its variants exhibit\nimmense potential in tackling combinatorial optimization challenges. However,\ntheir practical realization confronts a dilemma: the requisite circuit depth\nfor satisfactory performance is problem-specific and often exceeds the maximum\ncapability of current quantum devices. To address this dilemma, here we first\nanalyze the convergence behavior of QAOA, uncovering the origins of this\ndilemma and elucidating the intricate relationship between the employed mixer\nHamiltonian, the specific problem at hand, and the permissible maximum circuit\ndepth. Harnessing this understanding, we introduce the Mixer Generator Network\n(MG-Net), a unified deep learning framework adept at dynamically formulating\noptimal mixer Hamiltonians tailored to distinct tasks and circuit depths.\nSystematic simulations, encompassing Ising models and weighted Max-Cut\ninstances with up to 64 qubits, substantiate our theoretical findings,\nhighlighting MG-Net's superior performance in terms of both approximation ratio\nand efficiency.\n", "date": "2024-09-27", "categories": []} {"id": "2403.08757", "title": "Efficient Combinatorial Optimization via Heat Diffusion", "abstract": " Combinatorial optimization problems are widespread but inherently challenging\ndue to their discrete nature. The primary limitation of existing methods is\nthat they can only access a small fraction of the solution space at each\niteration, resulting in limited efficiency for searching the global optimal. To\novercome this challenge, diverging from conventional efforts of expanding the\nsolver's search scope, we focus on enabling information to actively propagate\nto the solver through heat diffusion. By transforming the target function while\npreserving its optima, heat diffusion facilitates information flow from distant\nregions to the solver, providing more efficient navigation. Utilizing heat\ndiffusion, we propose a framework for solving general combinatorial\noptimization problems. The proposed methodology demonstrates superior\nperformance across a range of the most challenging and widely encountered\ncombinatorial optimizations. Echoing recent advancements in harnessing\nthermodynamics for generative artificial intelligence, our study further\nreveals its significant potential in advancing combinatorial optimization.\n", "date": "2024-03-13", "categories": []} {"id": "2311.00405", "title": "Couples can be tractable: New algorithms and hardness results for the\n Hospitals / Residents problem with Couples", "abstract": " In this paper, we study the Hospitals / Residents problem with Couples (HRC),\nwhere a solution is a stable matching or a report that none exists. We present\na novel polynomial-time algorithm that can find a near-feasible stable matching\n(adjusting the hospitals' capacities by at most 1) in an HRC instance where the\ncouples' preferences are sub-responsive (i.e., if one member switches to a\nbetter hospital, then the couple also improves) and sub-complete (i.e., each\npair of hospitals that are individually acceptable to both members are jointly\nacceptable for the couple) by reducing it to an instance of the Stable Fixtures\nproblem. We also present a polynomial-time algorithm for HRC in a\nsub-responsive, sub-complete instance that is a Dual Market, or where all\ncouples are one of several possible types. We show that our algorithm also\nimplies the polynomial-time solvability of a stable b-matching problem, where\nthe underlying graph is a multigraph with loops.\n We complement our algorithms with several hardness results. We show that HRC\nwith sub-responsive and sub-complete couples is NP-hard, even with other strong\nrestrictions. We also show that HRC with a Dual Market is NP-hard under several\nsimultaneous restrictions. Finally, we show that the problem of finding a\nmatching with the minimum number of blocking pairs in HRC is not approximable\nwithin $m^{1-\\varepsilon}$, for any $\\varepsilon>0$, where $m$ is the total\nlength of the hospitals' preference lists, unless P=NP, even if each couple\napplies to only one pair of hospitals.\n Our polynomial-time solvability results greatly expand the class of known\ntractable instances of HRC and provide a useful tool for designing better and\nmore efficient mechanisms in the future.\n", "date": "2023-11-01", "categories": []} {"id": "2404.13737", "title": "Stochastic Multi-round Submodular Optimization with Budget", "abstract": " In this work, we study the Stochastic Budgeted Multi-round Submodular\nMaximization (SBMSm) problem, where we aim to adaptively maximize the sum, over\nmultiple rounds, of a monotone and submodular objective function defined on\nsubsets of items. The objective function also depends on the realization of\nstochastic events, and the total number of items we can select over all rounds\nis bounded by a limited budget. This problem extends, and generalizes to\nmultiple round settings, well-studied problems such as (adaptive) influence\nmaximization and stochastic probing.\n We show that, if the number of items and stochastic events is somehow\nbounded, there is a polynomial time dynamic programming algorithm for SBMSm.\nThen, we provide a simple greedy $1/2(1-1/e-\\epsilon)\\approx\n0.316$-approximation algorithm for SBMSm, that first non-adaptively allocates\nthe budget to be spent at each round, and then greedily and adaptively\nmaximizes the objective function by using the budget assigned at each round.\nFinally, we introduce the {\\em budget-adaptivity gap}, by which we measure how\nmuch an adaptive policy for SBMSm is better than an optimal partially adaptive\none that, as in our greedy algorithm, determines the budget allocation in\nadvance. We show that the budget-adaptivity gap lies between $e/(e-1)\\approx\n1.582$ and $2$.\n", "date": "2024-04-21", "categories": []} {"id": "2409.09677", "title": "Mitigating Dimensionality in 2D Rectangle Packing Problem under\n Reinforcement Learning Schema", "abstract": " This paper explores the application of Reinforcement Learning (RL) to the\ntwo-dimensional rectangular packing problem. We propose a reduced\nrepresentation of the state and action spaces that allow us for high\ngranularity. Leveraging UNet architecture and Proximal Policy Optimization\n(PPO), we achieved a model that is comparable to the MaxRect heuristic.\nHowever, our approach has great potential to be generalized to nonrectangular\npacking problems and complex constraints.\n", "date": "2024-09-15", "categories": []} {"id": "2409.14173", "title": "An Evolutionary Algorithm For the Vehicle Routing Problem with Drones\n with Interceptions", "abstract": " The use of trucks and drones as a solution to address last-mile delivery\nchallenges is a new and promising research direction explored in this paper.\nThe variation of the problem where the drone can intercept the truck while in\nmovement or at the customer location is part of an optimisation problem called\nthe vehicle routing problem with drones with interception (VRPDi). This paper\nproposes an evolutionary algorithm to solve the VRPDi. In this variation of the\nVRPDi, multiple pairs of trucks and drones need to be scheduled. The pairs\nleave and return to a depot location together or separately to make deliveries\nto customer nodes. The drone can intercept the truck after the delivery or meet\nup with the truck at the following customer location. The algorithm was\nexecuted on the travelling salesman problem with drones (TSPD) datasets by\nBouman et al. (2015), and the performance of the algorithm was compared by\nbenchmarking the results of the VRPDi against the results of the VRP of the\nsame dataset. This comparison showed improvements in total delivery time\nbetween 39% and 60%. Further detailed analysis of the algorithm results\nexamined the total delivery time, distance, node delivery scheduling and the\ndegree of diversity during the algorithm execution. This analysis also\nconsidered how the algorithm handled the VRPDi constraints. The results of the\nalgorithm were then benchmarked against algorithms in Dillon et al. (2023) and\nErnst (2024). The latter solved the problem with a maximum drone distance\nconstraint added to the VRPDi. The analysis and benchmarking of the algorithm\nresults showed that the algorithm satisfactorily solved 50 and 100-nodes\nproblems in a reasonable amount of time, and the solutions found were better\nthan those found by the algorithms in Dillon et al. (2023) and Ernst (2024) for\nthe same problems.\n", "date": "2024-09-21", "categories": []} {"id": "2409.04464", "title": "Leveraging Large Language Models for Solving Rare MIP Challenges", "abstract": " Mixed Integer Programming (MIP) has been extensively applied in areas\nrequiring mathematical solvers to address complex instances within tight time\nconstraints. However, as the problem scale increases, the complexity of model\nformulation and finding feasible solutions escalates significantly. In\ncontrast, the model-building cost for end-to-end models, such as large language\nmodels (LLMs), remains largely unaffected by problem scale due to their pattern\nrecognition capabilities. While LLMs, like GPT-4, without fine-tuning, can\nhandle some traditional medium-scale MIP problems, they struggle with uncommon\nor highly specialized MIP scenarios. Fine-tuning LLMs can yield some feasible\nsolutions for medium-scale MIP instances, but these models typically fail to\nexplore diverse solutions when constrained by a low and constant temperature,\nlimiting their performance. In this paper, we propose and evaluate a\nrecursively dynamic temperature method integrated with a chain-of-thought\napproach. Our findings show that starting with a high temperature and gradually\nlowering it leads to better feasible solutions compared to other dynamic\ntemperature strategies. Additionally, by comparing results generated by the LLM\nwith those from Gurobi, we demonstrate that the LLM can produce solutions that\ncomplement traditional solvers by accelerating the pruning process and\nimproving overall efficiency.\n", "date": "2024-09-03", "categories": []} {"id": "2409.11933", "title": "Reinforcement Learning as an Improvement Heuristic for Real-World\n Production Scheduling", "abstract": " The integration of Reinforcement Learning (RL) with heuristic methods is an\nemerging trend for solving optimization problems, which leverages RL's ability\nto learn from the data generated during the search process. One promising\napproach is to train an RL agent as an improvement heuristic, starting with a\nsuboptimal solution that is iteratively improved by applying small changes. We\napply this approach to a real-world multiobjective production scheduling\nproblem. Our approach utilizes a network architecture that includes Transformer\nencoding to learn the relationships between jobs. Afterwards, a probability\nmatrix is generated from which pairs of jobs are sampled and then swapped to\nimprove the solution. We benchmarked our approach against other heuristics\nusing real data from our industry partner, demonstrating its superior\nperformance.\n", "date": "2024-09-18", "categories": []} {"id": "2409.10680", "title": "Multi-agent Path Finding in Continuous Environment", "abstract": " We address a variant of multi-agent path finding in continuous environment\n(CE-MAPF), where agents move along sets of smooth curves. Collisions between\nagents are resolved via avoidance in the space domain. A new Continuous\nEnvironment Conflict-Based Search (CE-CBS) algorithm is proposed in this work.\nCE-CBS combines conflict-based search (CBS) for the high-level search framework\nwith RRT* for low-level path planning. The CE-CBS algorithm is tested under\nvarious settings on diverse CE-MAPF instances. Experimental results show that\nCE-CBS is competitive w.r.t. to other algorithms that consider continuous\naspect in MAPF such as MAPF with continuous time.\n", "date": "2024-09-16", "categories": []} {"id": "2409.11091", "title": "Online Combinatorial Allocations and Auctions with Few Samples", "abstract": " In online combinatorial allocations/auctions, n bidders sequentially arrive,\neach with a combinatorial valuation (such as submodular/XOS) over subsets of m\nindivisible items. The aim is to immediately allocate a subset of the remaining\nitems to maximize the total welfare, defined as the sum of bidder valuations. A\nlong line of work has studied this problem when the bidder valuations come from\nknown independent distributions. In particular, for submodular/XOS valuations,\nwe know 2-competitive algorithms/mechanisms that set a fixed price for each\nitem and the arriving bidders take their favorite subset of the remaining items\ngiven these prices. However, these algorithms traditionally presume the\navailability of the underlying distributions as part of the input to the\nalgorithm. Contrary to this assumption, practical scenarios often require the\nlearning of distributions, a task complicated by limited sample availability.\nThis paper investigates the feasibility of achieving O(1)-competitive\nalgorithms under the realistic constraint of having access to only a limited\nnumber of samples from the underlying bidder distributions.\n Our first main contribution shows that a mere single sample from each bidder\ndistribution is sufficient to yield an O(1)-competitive algorithm for\nsubmodular/XOS valuations. This result leverages a novel extension of the\nsecretary-style analysis, employing the sample to have the algorithm compete\nagainst itself. Although online, this first approach does not provide an online\ntruthful mechanism. Our second main contribution shows that a polynomial number\nof samples suffices to yield a $(2+\\epsilon)$-competitive online truthful\nmechanism for submodular/XOS valuations and any constant $\\epsilon>0$. This\nresult is based on a generalization of the median-based algorithm for the\nsingle-item prophet inequality problem to combinatorial settings with multiple\nitems.\n", "date": "2024-09-17", "categories": []} {"id": "2409.11290", "title": "Neural Networks for Vehicle Routing Problem", "abstract": " The Vehicle Routing Problem is about optimizing the routes of vehicles to\nmeet the needs of customers at specific locations. The route graph consists of\ndepots on several levels and customer positions. Several optimization methods\nhave been developed over the years, most of which are based on some type of\nclassic heuristic: genetic algorithm, simulated annealing, tabu search, ant\ncolony optimization, firefly algorithm. Recent developments in machine learning\nprovide a new toolset, the rich family of neural networks, for tackling complex\nproblems. The main area of application of neural networks is the area of\nclassification and regression. Route optimization can be viewed as a new\nchallenge for neural networks. The article first presents an analysis of the\napplicability of neural network tools, then a novel graphical neural network\nmodel is presented in detail. The efficiency analysis based on test experiments\nshows the applicability of the proposed NN architecture.\n", "date": "2024-09-17", "categories": []} {"id": "2206.09563", "title": "Scalable Distributed Algorithms for Size-Constrained Submodular\n Maximization in the MapReduce and Adaptive Complexity Models", "abstract": " Distributed maximization of a submodular function in the MapReduce (MR) model\nhas received much attention, culminating in two frameworks that allow a\ncentralized algorithm to be run in the MR setting without loss of\napproximation, as long as the centralized algorithm satisfies a certain\nconsistency property -- which had previously only been known to be satisfied by\nthe standard greedy and continous greedy algorithms. A separate line of work\nhas studied parallelizability of submodular maximization in the adaptive\ncomplexity model, where each thread may have access to the entire ground set.\nFor the size-constrained maximization of a monotone and submodular function, we\nshow that several sublinearly adaptive (highly parallelizable) algorithms\nsatisfy the consistency property required to work in the MR setting, which\nyields practical, parallelizable and distributed algorithms. Separately, we\ndevelop the first distributed algorithm with linear query complexity for this\nproblem. Finally, we provide a method to increase the maximum cardinality\nconstraint for MR algorithms at the cost of additional MR rounds.\n", "date": "2022-06-20", "categories": []} {"id": "2209.01535", "title": "Simulation-Assisted Optimization for Large-Scale Evacuation Planning\n with Congestion-Dependent Delays", "abstract": " Evacuation planning is a crucial part of disaster management. However, joint\noptimization of its two essential components, routing and scheduling, with\nobjectives such as minimizing average evacuation time or evacuation completion\ntime, is a computationally hard problem. To approach it, we present MIP-LNS, a\nscalable optimization method that utilizes heuristic search with mathematical\noptimization and can optimize a variety of objective functions. We also present\nthe method MIP-LNS-SIM, where we combine agent-based simulation with MIP-LNS to\nestimate delays due to congestion, as well as, find optimized plans considering\nsuch delays. We use Harris County in Houston, Texas, as our study area. We show\nthat, within a given time limit, MIP-LNS finds better solutions than existing\nmethods in terms of three different metrics. However, when congestion dependent\ndelay is considered, MIP-LNS-SIM outperforms MIP-LNS in multiple performance\nmetrics. In addition, MIP-LNS-SIM has a significantly lower percent error in\nestimated evacuation completion time compared to MIP-LNS.\n", "date": "2022-09-04", "categories": []} {"id": "2409.07547", "title": "Machine Learning and Constraint Programming for Efficient Healthcare\n Scheduling", "abstract": " Solving combinatorial optimization problems involve satisfying a set of hard\nconstraints while optimizing some objectives. In this context, exact or\napproximate methods can be used. While exact methods guarantee the optimal\nsolution, they often come with an exponential running time as opposed to\napproximate methods that trade the solutions quality for a better running time.\nIn this context, we tackle the Nurse Scheduling Problem (NSP). The NSP consist\nin assigning nurses to daily shifts within a planning horizon such that\nworkload constraints are satisfied while hospitals costs and nurses preferences\nare optimized. To solve the NSP, we propose implicit and explicit approaches.\nIn the implicit solving approach, we rely on Machine Learning methods using\nhistorical data to learn and generate new solutions through the constraints and\nobjectives that may be embedded in the learned patterns. To quantify the\nquality of using our implicit approach in capturing the embedded constraints\nand objectives, we rely on the Frobenius Norm, a quality measure used to\ncompute the average error between the generated solutions and historical data.\nTo compensate for the uncertainty related to the implicit approach given that\nthe constraints and objectives may not be concretely visible in the produced\nsolutions, we propose an alternative explicit approach where we first model the\nNSP using the Constraint Satisfaction Problem (CSP) framework. Then we develop\nStochastic Local Search methods and a new Branch and Bound algorithm enhanced\nwith constraint propagation techniques and variables/values ordering\nheuristics. Since our implicit approach may not guarantee the feasibility or\noptimality of the generated solution, we propose a data-driven approach to\npassively learn the NSP as a constraint network. The learned constraint\nnetwork, formulated as a CSP, will then be solved using the methods we listed\nearlier.\n", "date": "2024-09-11", "categories": []} {"id": "2409.05888", "title": "MA-CDMR: An Intelligent Cross-domain Multicast Routing Method based on\n Multiagent Deep Reinforcement Learning in Multi-domain SDWN", "abstract": " The cross-domain multicast routing problem in a software-defined wireless\nnetwork with multiple controllers is a classic NP-hard optimization problem. As\nthe network size increases, designing and implementing cross-domain multicast\nrouting paths in the network requires not only designing efficient solution\nalgorithms to obtain the optimal cross-domain multicast tree but also ensuring\nthe timely and flexible acquisition and maintenance of global network state\ninformation. However, existing solutions have a limited ability to sense the\nnetwork traffic state, affecting the quality of service of multicast services.\nIn addition, these methods have difficulty adapting to the highly dynamically\nchanging network states and have slow convergence speeds. To this end, this\npaper aims to design and implement a multiagent deep reinforcement learning\nbased cross-domain multicast routing method for SDWN with multicontroller\ndomains. First, a multicontroller communication mechanism and a multicast group\nmanagement module are designed to transfer and synchronize network information\nbetween different control domains of the SDWN, thus effectively managing the\njoining and classification of members in the cross-domain multicast group.\nSecond, a theoretical analysis and proof show that the optimal cross-domain\nmulticast tree includes an interdomain multicast tree and an intradomain\nmulticast tree. An agent is established for each controller, and a cooperation\nmechanism between multiple agents is designed to effectively optimize\ncross-domain multicast routing and ensure consistency and validity in the\nrepresentation of network state information for cross-domain multicast routing\ndecisions. Third, a multiagent reinforcement learning-based method that\ncombines online and offline training is designed to reduce the dependence on\nthe real-time environment and increase the convergence speed of multiple\nagents.\n", "date": "2024-08-27", "categories": []} {"id": "2409.04456", "title": "Pattern based learning and optimisation through pricing for bin packing\n problem", "abstract": " As a popular form of knowledge and experience, patterns and their\nidentification have been critical tasks in most data mining applications.\nHowever, as far as we are aware, no study has systematically examined the\ndynamics of pattern values and their reuse under varying conditions. We argue\nthat when problem conditions such as the distributions of random variables\nchange, the patterns that performed well in previous circumstances may become\nless effective and adoption of these patterns would result in sub-optimal\nsolutions. In response, we make a connection between data mining and the\nduality theory in operations research and propose a novel scheme to efficiently\nidentify patterns and dynamically quantify their values for each specific\ncondition. Our method quantifies the value of patterns based on their ability\nto satisfy stochastic constraints and their effects on the objective value,\nallowing high-quality patterns and their combinations to be detected. We use\nthe online bin packing problem to evaluate the effectiveness of the proposed\nscheme and illustrate the online packing procedure with the guidance of\npatterns that address the inherent uncertainty of the problem. Results show\nthat the proposed algorithm significantly outperforms the state-of-the-art\nmethods. We also analysed in detail the distinctive features of the proposed\nmethods that lead to performance improvement and the special cases where our\nmethod can be further improved.\n", "date": "2024-08-27", "categories": []} {"id": "2409.04926", "title": "A $\\Delta$-evaluation function for column permutation problems", "abstract": " In this study, a new $\\Delta$-evaluation method is introduced for solving a\ncolumn permutation problem defined on a sparse binary matrix with the\nconsecutive ones property. This problem models various $\\mathcal{NP}$-hard\nproblems in graph theory and industrial manufacturing contexts. The\ncomputational experiments compare the processing time of the\n$\\Delta$-evaluation method with two other methods used in well-known local\nsearch procedures. The study considers a comprehensive set of instances of\nwell-known problems, such as Gate Matrix Layout and Minimization of Open\nStacks. The proposed evaluation method is generally competitive and\nparticularly useful for large and dense instances. It can be easily integrated\ninto local search and metaheuristic algorithms to improve solutions without\nsignificantly increasing processing time.\n", "date": "2024-09-07", "categories": []} {"id": "2409.05265", "title": "Learning Submodular Sequencing from Samples", "abstract": " This paper addresses the problem of sequential submodular maximization:\nselecting and ranking items in a sequence to optimize some composite submodular\nfunction. In contrast to most of the previous works, which assume access to the\nutility function, we assume that we are given only a set of samples. Each\nsample includes a random sequence of items and its associated utility. We\npresent an algorithm that, given polynomially many samples drawn from a\ntwo-stage uniform distribution, achieves an approximation ratio dependent on\nthe curvature of individual submodular functions. Our results apply in a wide\nvariety of real-world scenarios, such as ranking products in online retail\nplatforms, where complete knowledge of the utility function is often impossible\nto obtain. Our algorithm gives an empirically useful solution in such contexts,\nthus proving that limited data can be of great use in sequencing tasks. From a\ntechnical perspective, our results extend prior work on ``optimization from\nsamples'' by generalizing from optimizing a set function to a\nsequence-dependent function.\n", "date": "2024-09-09", "categories": []} {"id": "2409.01588", "title": "Large-scale Urban Facility Location Selection with Knowledge-informed\n Reinforcement Learning", "abstract": " The facility location problem (FLP) is a classical combinatorial optimization\nchallenge aimed at strategically laying out facilities to maximize their\naccessibility. In this paper, we propose a reinforcement learning method\ntailored to solve large-scale urban FLP, capable of producing near-optimal\nsolutions at superfast inference speed. We distill the essential swap operation\nfrom local search, and simulate it by intelligently selecting edges on a graph\nof urban regions, guided by a knowledge-informed graph neural network, thus\nsidestepping the need for heavy computation of local search. Extensive\nexperiments on four US cities with different geospatial conditions demonstrate\nthat our approach can achieve comparable performance to commercial solvers with\nless than 5\\% accessibility loss, while displaying up to 1000 times speedup. We\ndeploy our model as an online geospatial application at\nhttps://huggingface.co/spaces/randommmm/MFLP.\n", "date": "2024-09-03", "categories": []} {"id": "2409.03881", "title": "Multi-agent Path Finding for Mixed Autonomy Traffic Coordination", "abstract": " In the evolving landscape of urban mobility, the prospective integration of\nConnected and Automated Vehicles (CAVs) with Human-Driven Vehicles (HDVs)\npresents a complex array of challenges and opportunities for autonomous driving\nsystems. While recent advancements in robotics have yielded Multi-Agent Path\nFinding (MAPF) algorithms tailored for agent coordination task characterized by\nsimplified kinematics and complete control over agent behaviors, these\nsolutions are inapplicable in mixed-traffic environments where uncontrollable\nHDVs must coexist and interact with CAVs. Addressing this gap, we propose the\nBehavior Prediction Kinematic Priority Based Search (BK-PBS), which leverages\nan offline-trained conditional prediction model to forecast HDV responses to\nCAV maneuvers, integrating these insights into a Priority Based Search (PBS)\nwhere the A* search proceeds over motion primitives to accommodate kinematic\nconstraints. We compare BK-PBS with CAV planning algorithms derived by\nrule-based car-following models, and reinforcement learning. Through\ncomprehensive simulation on a highway merging scenario across diverse scenarios\nof CAV penetration rate and traffic density, BK-PBS outperforms these baselines\nin reducing collision rates and enhancing system-level travel delay. Our work\nis directly applicable to many scenarios of multi-human multi-robot\ncoordination.\n", "date": "2024-09-05", "categories": []} {"id": "2409.04415", "title": "Improved Parallel Algorithm for Non-Monotone Submodular Maximization\n under Knapsack Constraint", "abstract": " This work proposes an efficient parallel algorithm for non-monotone\nsubmodular maximization under a knapsack constraint problem over the ground set\nof size $n$. Our algorithm improves the best approximation factor of the\nexisting parallel one from $8+\\epsilon$ to $7+\\epsilon$ with $O(\\log n)$\nadaptive complexity.\n The key idea of our approach is to create a new alternate threshold\nalgorithmic framework. This strategy alternately constructs two disjoint\ncandidate solutions within a constant number of sequence rounds. Then, the\nalgorithm boosts solution quality without sacrificing the adaptive complexity.\nExtensive experimental studies on three applications, Revenue Maximization,\nImage Summarization, and Maximum Weighted Cut, show that our algorithm not only\nsignificantly increases solution quality but also requires comparative\nadaptivity to state-of-the-art algorithms.\n", "date": "2024-09-06", "categories": []} {"id": "2409.02943", "title": "A Note On Deterministic Submodular Maximization With Bounded Curvature", "abstract": " We show that the recent breakthrough result of [Buchbinder and Feldman,\nFOCS'24] could further lead to a deterministic\n$(1-\\kappa_{f}/e-\\varepsilon)$-approximate algorithm for maximizing a\nsubmodular function with curvature $\\kappa_{f}$ under matroid constraint.\n", "date": "2024-08-25", "categories": []} {"id": "2409.03545", "title": "The Power of Second Chance: Personalized Submodular Maximization with\n Two Candidates", "abstract": " Most of existing studies on submodular maximization focus on selecting a\nsubset of items that maximizes a \\emph{single} submodular function. However, in\nmany real-world scenarios, we might have multiple user-specific functions, each\nof which models the utility of a particular type of user. In these settings,\nour goal would be to choose a set of items that performs well across all the\nuser-specific functions. One way to tackle this problem is to select a single\nsubset that maximizes the sum of all of the user-specific functions. Although\nthis aggregate approach is efficient in the sense that it avoids computation of\nsets for individual functions, it really misses the power of personalization -\nfor it does not allow to choose different sets for different functions. In this\npaper, we introduce the problem of personalized submodular maximization with\ntwo candidate solutions. For any two candidate solutions, the utility of each\nuser-specific function is defined as the better of these two candidates. Our\nobjective is, therefore, to select the best set of two candidates that maximize\nthe sum of utilities of all the user-specific functions. We have designed\neffective algorithms for this problem. We also discuss how our approach\ngeneralizes to multiple candidate solutions, increasing flexibility and\npersonalization in our solution.\n", "date": "2024-09-05", "categories": []} {"id": "2302.05405", "title": "ACE, a generic constraint solver", "abstract": " Constraint Programming (CP) is a useful technology for modeling and solving\ncombinatorial constrained problems. On the one hand, on can use a library like\nPyCSP3 for easily modeling problems arising in various application fields\n(e.g., scheduling, planning, data-mining, cryptography, bio-informatics,\norganic chemistry, etc.). Problem instances can then be directly generated from\nspecific models and data. On the other hand, for solving instances (notably,\nrepresented in XCSP3 format), one can use a constraint solver like ACE, which\nis presented in this paper. ACE is an open-source constraint solver, developed\nin Java, which focuses on integer variables (including 0/1-Boolean variables),\nstate-of-the-art table constraints, popular global constraints, search\nheuristics and (mono-criterion) optimization.\n", "date": "2023-01-06", "categories": []} {"id": "2409.00670", "title": "Towards Faster Graph Partitioning via Pre-training and Inductive\n Inference", "abstract": " Graph partitioning (GP) is a classic problem that divides the node set of a\ngraph into densely-connected blocks. Following the IEEE HPEC Graph Challenge\nand recent advances in pre-training techniques (e.g., large-language models),\nwe propose PR-GPT (Pre-trained & Refined Graph ParTitioning) based on a novel\npre-training & refinement paradigm. We first conduct the offline pre-training\nof a deep graph learning (DGL) model on small synthetic graphs with various\ntopology properties. By using the inductive inference of DGL, one can directly\ngeneralize the pre-trained model (with frozen model parameters) to large graphs\nand derive feasible GP results. We also use the derived partition as a good\ninitialization of an efficient GP method (e.g., InfoMap) to further refine the\nquality of partitioning. In this setting, the online generalization and\nrefinement of PR-GPT can not only benefit from the transfer ability regarding\nquality but also ensure high inference efficiency without re-training. Based on\na mechanism of reducing the scale of a graph to be processed by the refinement\nmethod, PR-GPT also has the potential to support streaming GP. Experiments on\nthe Graph Challenge benchmark demonstrate that PR-GPT can ensure faster GP on\nlarge-scale graphs without significant quality degradation, compared with\nrunning a refinement method from scratch. We will make our code public at\nhttps://github.com/KuroginQin/PRGPT.\n", "date": "2024-09-01", "categories": []} {"id": "2409.00968", "title": "Solving Integrated Process Planning and Scheduling Problem via Graph\n Neural Network Based Deep Reinforcement Learning", "abstract": " The Integrated Process Planning and Scheduling (IPPS) problem combines\nprocess route planning and shop scheduling to achieve high efficiency in\nmanufacturing and maximize resource utilization, which is crucial for modern\nmanufacturing systems. Traditional methods using Mixed Integer Linear\nProgramming (MILP) and heuristic algorithms can not well balance solution\nquality and speed when solving IPPS. In this paper, we propose a novel\nend-to-end Deep Reinforcement Learning (DRL) method. We model the IPPS problem\nas a Markov Decision Process (MDP) and employ a Heterogeneous Graph Neural\nNetwork (GNN) to capture the complex relationships among operations, machines,\nand jobs. To optimize the scheduling strategy, we use Proximal Policy\nOptimization (PPO). Experimental results show that, compared to traditional\nmethods, our approach significantly improves solution efficiency and quality in\nlarge-scale IPPS instances, providing superior scheduling strategies for modern\nintelligent manufacturing systems.\n", "date": "2024-09-02", "categories": []} {"id": "2409.01815", "title": "Learning State-Dependent Policy Parametrizations for Dynamic Technician\n Routing with Rework", "abstract": " Home repair and installation services require technicians to visit customers\nand resolve tasks of different complexity. Technicians often have heterogeneous\nskills and working experiences. The geographical spread of customers makes\nachieving only perfect matches between technician skills and task requirements\nimpractical. Additionally, technicians are regularly absent due to sickness.\nWith non-perfect assignments regarding task requirement and technician skill,\nsome tasks may remain unresolved and require a revisit and rework. Companies\nseek to minimize customer inconvenience due to delay. We model the problem as a\nsequential decision process where, over a number of service days, customers\nrequest service while heterogeneously skilled technicians are routed to serve\ncustomers in the system. Each day, our policy iteratively builds tours by\nadding \"important\" customers. The importance bases on analytical considerations\nand is measured by respecting routing efficiency, urgency of service, and risk\nof rework in an integrated fashion. We propose a state-dependent balance of\nthese factors via reinforcement learning. A comprehensive study shows that\ntaking a few non-perfect assignments can be quite beneficial for the overall\nservice quality. We further demonstrate the value provided by a state-dependent\nparametrization.\n", "date": "2024-09-03", "categories": []} {"id": "2404.16379", "title": "Optimal and Bounded Suboptimal Any-Angle Multi-agent Pathfinding", "abstract": " Multi-agent pathfinding (MAPF) is the problem of finding a set of\nconflict-free paths for a set of agents. Typically, the agents' moves are\nlimited to a pre-defined graph of possible locations and allowed transitions\nbetween them, e.g. a 4-neighborhood grid. We explore how to solve MAPF problems\nwhen each agent can move between any pair of possible locations as long as\ntraversing the line segment connecting them does not lead to a collision with\nthe obstacles. This is known as any-angle pathfinding. We present the first\noptimal any-angle multi-agent pathfinding algorithm. Our planner is based on\nthe Continuous Conflict-based Search (CCBS) algorithm and an optimal any-angle\nvariant of the Safe Interval Path Planning (TO-AA-SIPP). The straightforward\ncombination of those, however, scales poorly since any-angle path finding\ninduces search trees with a very large branching factor. To mitigate this, we\nadapt two techniques from classical MAPF to the any-angle setting, namely\nDisjoint Splitting and Multi-Constraints. Experimental results on different\ncombinations of these techniques show they enable solving over 30% more\nproblems than the vanilla combination of CCBS and TO-AA-SIPP. In addition, we\npresent a bounded-suboptimal variant of our algorithm, that enables trading\nruntime for solution cost in a controlled manner.\n", "date": "2024-04-25", "categories": []} {"id": "2408.16890", "title": "Robotic warehousing operations: a learn-then-optimize approach to\n large-scale neighborhood search", "abstract": " The rapid deployment of robotics technologies requires dedicated optimization\nalgorithms to manage large fleets of autonomous agents. This paper supports\nrobotic parts-to-picker operations in warehousing by optimizing\norder-workstation assignments, item-pod assignments and the schedule of order\nfulfillment at workstations. The model maximizes throughput, while managing\nhuman workload at the workstations and congestion in the facility. We solve it\nvia large-scale neighborhood search, with a novel learn-then-optimize approach\nto subproblem generation. The algorithm relies on an offline machine learning\nprocedure to predict objective improvements based on subproblem features, and\nan online optimization model to generate a new subproblem at each iteration. In\ncollaboration with Amazon Robotics, we show that our model and algorithm\ngenerate much stronger solutions for practical problems than state-of-the-art\napproaches. In particular, our solution enhances the utilization of robotic\nfleets by coordinating robotic tasks for human operators to pick multiple items\nat once, and by coordinating robotic routes to avoid congestion in the\nfacility.\n", "date": "2024-08-29", "categories": []} {"id": "2408.14864", "title": "Dynamic operator management in meta-heuristics using reinforcement\n learning: an application to permutation flowshop scheduling problems", "abstract": " This study develops a framework based on reinforcement learning to\ndynamically manage a large portfolio of search operators within\nmeta-heuristics. Using the idea of tabu search, the framework allows for\ncontinuous adaptation by temporarily excluding less efficient operators and\nupdating the portfolio composition during the search. A Q-learning-based\nadaptive operator selection mechanism is used to select the most suitable\noperator from the dynamically updated portfolio at each stage. Unlike\ntraditional approaches, the proposed framework requires no input from the\nexperts regarding the search operators, allowing domain-specific non-experts to\neffectively use the framework. The performance of the proposed framework is\nanalyzed through an application to the permutation flowshop scheduling problem.\nThe results demonstrate the superior performance of the proposed framework\nagainst state-of-the-art algorithms in terms of optimality gap and convergence\nspeed.\n", "date": "2024-08-27", "categories": []} {"id": "2405.00285", "title": "iMTSP: Solving Min-Max Multiple Traveling Salesman Problem with\n Imperative Learning", "abstract": " This paper considers a Min-Max Multiple Traveling Salesman Problem (MTSP),\nwhere the goal is to find a set of tours, one for each agent, to collectively\nvisit all the cities while minimizing the length of the longest tour. Though\nMTSP has been widely studied, obtaining near-optimal solutions for large-scale\nproblems is still challenging due to its NP-hardness. Recent efforts in\ndata-driven methods face challenges of the need for hard-to-obtain supervision\nand issues with high variance in gradient estimations, leading to slow\nconvergence and highly suboptimal solutions. We address these issues by\nreformulating MTSP as a bilevel optimization problem, using the concept of\nimperative learning (IL). This involves introducing an allocation network that\ndecomposes the MTSP into multiple single-agent traveling salesman problems\n(TSPs). The longest tour from these TSP solutions is then used to\nself-supervise the allocation network, resulting in a new self-supervised,\nbilevel, end-to-end learning framework, which we refer to as imperative MTSP\n(iMTSP). Additionally, to tackle the high-variance gradient issues during the\noptimization, we introduce a control variate-based gradient estimation\nalgorithm. Our experiments showed that these innovative designs enable our\ngradient estimator to converge 20% faster than the advanced reinforcement\nlearning baseline and find up to 80% shorter tour length compared with Google\nOR-Tools MTSP solver, especially in large-scale problems (e.g. 1000 cities and\n15 agents).\n", "date": "2024-05-01", "categories": []} {"id": "2408.12032", "title": "A Constraint Programming Approach to Fair High School Course Scheduling", "abstract": " Issues of inequity in U.S. high schools' course scheduling did not previously\nexist. However, in recent years, with the increase in student population and\ncourse variety, students perceive that the course scheduling method is unfair.\nCurrent integer programming (IP) methods to the high school scheduling problem\n(HSSP) fall short in addressing these fairness concerns. The purpose of this\nresearch is to develop a solution methodology that generates feasible and fair\ncourse schedules using student preferences. Utilizing principles of fairness,\nwhich have been well studied in market design, we define the fair high school\nscheduling problem (FHSSP), a novel extension to the HSSP, and devise a\ncorresponding algorithm based on integer programming to solve the FHSSP. We\ntest our approach on a real course request dataset from a high school in\nCalifornia, USA. Results show that our algorithm can generate schedules that\nare both feasible and fair. In this paper, we demonstrate that our IP algorithm\nnot only solves the HSSP and FHSSP in the United States but has the potential\nto be applied to various real-world scheduling problems. Additionally, we show\nthe feasibility of integrating human emotions into mathematical modeling.\n", "date": "2024-08-21", "categories": []} {"id": "2408.11187", "title": "Optimization of Multi-Agent Flying Sidekick Traveling Salesman Problem\n over Road Networks", "abstract": " The mixed truck-drone delivery systems have attracted increasing attention\nfor last-mile logistics, but real-world complexities demand a shift from\nsingle-agent, fully connected graph models to multi-agent systems operating on\nactual road networks. We introduce the multi-agent flying sidekick traveling\nsalesman problem (MA-FSTSP) on road networks, extending the single truck-drone\nmodel to multiple trucks, each carrying multiple drones while considering full\nroad networks for truck restrictions and flexible drone routes. We propose a\nmixed-integer linear programming model and an efficient three-phase heuristic\nalgorithm for this NP-hard problem. Our approach decomposes MA-FSTSP into\nmanageable subproblems of one truck with multiple drones. Then, it computes the\nroutes for trucks without drones in subproblems, which are used in the final\nphase as heuristics to help optimize drone and truck routes simultaneously.\nExtensive numerical experiments on Manhattan and Boston road networks\ndemonstrate our algorithm's superior effectiveness and efficiency,\nsignificantly outperforming both column generation and variable neighborhood\nsearch baselines in solution quality and computation time. Notably, our\napproach scales to more than 300 customers within a 5-minute time limit,\nshowcasing its potential for large-scale, real-world logistics applications.\n", "date": "2024-08-20", "categories": []} {"id": "2402.10725", "title": "Cloud Kitchen: Using Planning-based Composite AI to Optimize Food\n Delivery Processes", "abstract": " The global food delivery market provides many opportunities for AI-based\nservices that can improve the efficiency of feeding the world. This paper\npresents the Cloud Kitchen platform as a decision-making tool for restaurants\nwith food delivery and a simulator to evaluate the impact of the decisions. The\nplatform contains a Technology-Specific Bridge (TSB) that provides an interface\nfor communicating with restaurants or the simulator. TSB uses a planning domain\nmodel to represent decisions embedded in the Unified Planning Framework (UPF).\nDecision-making, which concerns allocating customers' orders to vehicles and\ndeciding in which order the customers will be served (for each vehicle), is\ndone via a Vehicle Routing Problem with Time Windows (VRPTW), an efficient tool\nfor this problem. We show that decisions made by our platform can improve\ncustomer satisfaction by reducing the number of delayed deliveries using a\nreal-world historical dataset.\n", "date": "2024-02-16", "categories": []} {"id": "2404.06492", "title": "Graph Reinforcement Learning for Combinatorial Optimization: A Survey\n and Unifying Perspective", "abstract": " Graphs are a natural representation for systems based on relations between\nconnected entities. Combinatorial optimization problems, which arise when\nconsidering an objective function related to a process of interest on discrete\nstructures, are often challenging due to the rapid growth of the solution\nspace. The trial-and-error paradigm of Reinforcement Learning has recently\nemerged as a promising alternative to traditional methods, such as exact\nalgorithms and (meta)heuristics, for discovering better decision-making\nstrategies in a variety of disciplines including chemistry, computer science,\nand statistics. Despite the fact that they arose in markedly different fields,\nthese techniques share significant commonalities. Therefore, we set out to\nsynthesize this work in a unifying perspective that we term Graph Reinforcement\nLearning, interpreting it as a constructive decision-making method for graph\nproblems. After covering the relevant technical background, we review works\nalong the dividing line of whether the goal is to optimize graph structure\ngiven a process of interest, or to optimize the outcome of the process itself\nunder fixed graph structure. Finally, we discuss the common challenges facing\nthe field and open research questions. In contrast with other surveys, the\npresent work focuses on non-canonical graph problems for which performant\nalgorithms are typically not known and Reinforcement Learning is able to\nprovide efficient and effective solutions.\n", "date": "2024-04-09", "categories": []} {"id": "2408.09028", "title": "On the Completeness of Conflict-Based Search: Temporally-Relative\n Duplicate Pruning", "abstract": " Conflict-Based Search (CBS) algorithm for the multi-agent pathfinding (MAPF)\nproblem is that it is incomplete for problems which have no solution; if no\nmitigating procedure is run in parallel, CBS will run forever when given an\nunsolvable problem instance. In this work, we introduce Temporally-Relative\nDuplicate Pruning (TRDP), a technique for duplicate detection and removal in\nboth classic and continuous-time MAPF domains. TRDP is a simple procedure which\ncloses the long-standing theoretic loophole of incompleteness for CBS by\ndetecting and avoiding the expansion of duplicate states. TRDP is shown both\ntheoretically and empirically to ensure termination without a significant\nimpact on runtime in the majority of problem instances. In certain cases, TRDP\nis shown to increase performance significantly\n", "date": "2024-08-16", "categories": []} {"id": "2408.09076", "title": "Twin Sorting Dynamic Programming Assisted User Association and Wireless\n Bandwidth Allocation for Hierarchical Federated Learning", "abstract": " In this paper, we study user association and wireless bandwidth allocation\nfor a hierarchical federated learning system that consists of mobile users,\nedge servers, and a cloud server. To minimize the length of a global round in\nhierarchical federated learning with equal bandwidth allocation, we formulate a\ncombinatorial optimization problem. We design the twin sorting dynamic\nprogramming (TSDP) algorithm that obtains a globally optimal solution in\npolynomial time when there are two edge servers. In addition, we put forward\nthe TSDP-assisted algorithm for user association when there are three or more\nedge servers. Furthermore, given a user association matrix, we formulate and\nsolve a convex optimization problem for optimal wireless bandwidth allocation.\nSimulation results show that the proposed approach outperforms a number of\nalternative schemes.\n", "date": "2024-08-17", "categories": []} {"id": "2408.10040", "title": "The Practimum-Optimum Algorithm for Manufacturing Scheduling: A Paradigm\n Shift Leading to Breakthroughs in Scale and Performance", "abstract": " The Practimum-Optimum (P-O) algorithm represents a paradigm shift in\ndeveloping automatic optimization products for complex real-life business\nproblems such as large-scale manufacturing scheduling. It leverages deep\nbusiness domain expertise to create a group of virtual human expert (VHE)\nagents with different \"schools of thought\" on how to create high-quality\nschedules. By computerizing them into algorithms, P-O generates many valid\nschedules at far higher speeds than human schedulers are capable of. Initially,\nthese schedules can also be local optimum peaks far away from high-quality\nschedules. By submitting these schedules to a reinforced machine learning\nalgorithm (RL), P-O learns the weaknesses and strengths of each VHE schedule,\nand accordingly derives reward and punishment changes in the Demand Set that\nwill modify the relative priorities for time and resource allocation that jobs\nreceived in the prior iteration that led to the current state of the schedule.\nThese cause the core logic of the VHE algorithms to explore, in the subsequent\niteration, substantially different parts of the schedules universe and\npotentially find higher-quality schedules. Using the hill climbing analogy,\nthis may be viewed as a big jump, shifting from a given local peak to a faraway\npromising start point equipped with knowledge embedded in the demand set for\nfuture iterations. This is a fundamental difference from most contemporary\nalgorithms, which spend considerable time on local micro-steps restricted to\nthe neighbourhoods of local peaks they visit. This difference enables a\nbreakthrough in scale and performance for fully automatic manufacturing\nscheduling in complex organizations. The P-O algorithm is at the heart of\nPlataine Scheduler that, in one click, routinely schedules 30,000-50,000 tasks\nfor real-life complex manufacturing operations.\n", "date": "2024-08-19", "categories": []} {"id": "2408.08484", "title": "An Unsupervised Learning Framework Combined with Heuristics for the\n Maximum Minimal Cut Problem", "abstract": " The Maximum Minimal Cut Problem (MMCP), a NP-hard combinatorial optimization\n(CO) problem, has not received much attention due to the demanding and\nchallenging bi-connectivity constraint. Moreover, as a CO problem, it is also a\ndaunting task for machine learning, especially without labeled instances. To\ndeal with these problems, this work proposes an unsupervised learning framework\ncombined with heuristics for MMCP that can provide valid and high-quality\nsolutions. As far as we know, this is the first work that explores machine\nlearning and heuristics to solve MMCP. The unsupervised solver is inspired by a\nrelaxation-plus-rounding approach, the relaxed solution is parameterized by\ngraph neural networks, and the cost and penalty of MMCP are explicitly written\nout, which can train the model end-to-end. A crucial observation is that each\nsolution corresponds to at least one spanning tree. Based on this finding, a\nheuristic solver that implements tree transformations by adding vertices is\nutilized to repair and improve the solution quality of the unsupervised solver.\nAlternatively, the graph is simplified while guaranteeing solution consistency,\nwhich reduces the running time. We conduct extensive experiments to evaluate\nour framework and give a specific application. The results demonstrate the\nsuperiority of our method against two techniques designed.\n", "date": "2024-08-16", "categories": []} {"id": "2408.08227", "title": "Evolving A* to Efficiently Solve the k Shortest-Path Problem (Extended\n Version)", "abstract": " The problem of finding the shortest path in a graph G(V, E) has been widely\nstudied. However, in many applications it is necessary to compute an arbitrary\nnumber of them, k. Even though the problem has raised a lot of interest from\ndifferent research communities and many applications of it are known, it has\nnot been addressed to the same extent as the single shortest path problem. The\nbest algorithm known for efficiently solving this task has a time complexity of\nO (|E| + |V|log{|V|}+k|V|)$ when computing paths in explicit form, and is based\non best-first search. This paper introduces a new search algorithm with the\nsame time complexity, which results from a natural evolution of A* thus, it\npreserves all its interesting properties, making it widely applicable to many\ndifferent domains. Experiments in various testbeds show a significant\nimprovement in performance over the state of the art, often by one or two\norders of magnitude.\n", "date": "2024-08-15", "categories": []} {"id": "2403.15961", "title": "SAT Encoding of Partial Ordering Models for Graph Coloring Problems", "abstract": " In this paper, we suggest new SAT encodings of the partial-ordering based ILP\nmodel for the graph coloring problem (GCP) and the bandwidth coloring problem\n(BCP). The GCP asks for the minimum number of colors that can be assigned to\nthe vertices of a given graph such that each two adjacent vertices get\ndifferent colors. The BCP is a generalization, where each edge has a weight\nthat enforces a minimal \"distance\" between the assigned colors, and the goal is\nto minimize the \"largest\" color used. For the widely studied GCP, we\nexperimentally compare our new SAT encoding to the state-of-the-art approaches\non the DIMACS benchmark set. Our evaluation confirms that this SAT encoding is\neffective for sparse graphs and even outperforms the state-of-the-art on some\nDIMACS instances. For the BCP, our theoretical analysis shows that the\npartial-ordering based SAT and ILP formulations have an asymptotically smaller\nsize than that of the classical assignment-based model. Our practical\nevaluation confirms not only a dominance compared to the assignment-based\nencodings but also to the state-of-the-art approaches on a set of benchmark\ninstances. Up to our knowledge, we have solved several open instances of the\nBCP from the literature for the first time.\n", "date": "2024-03-23", "categories": []} {"id": "2408.07417", "title": "The Restaurant Meal Delivery Problem with Ghost Kitchens", "abstract": " Restaurant meal delivery has been rapidly growing in the last few years. The\nmain challenges in operating it are the temporally and spatially dispersed\nstochastic demand that arrives from customers all over town as well as the\ncustomers' expectation of timely and fresh delivery. To overcome these\nchallenges a new business concept emerged, \"Ghost kitchens\". This concept\nproposes synchronized food preparation of several restaurants in a central\ncomplex, exploiting consolidation benefits. However, dynamically scheduling\nfood preparation and delivery is challenging and we propose operational\nstrategies for the effective operations of ghost kitchens. We model the problem\nas a sequential decision process. For the complex, combinatorial decision space\nof scheduling order preparations, consolidating orders to trips, and scheduling\ntrip departures, we propose a large neighborhood search procedure based on\npartial decisions and driven by analytical properties. Within the large\nneighborhood search, decisions are evaluated via a value function\napproximation, enabling anticipatory and real-time decision making. We show the\neffectiveness of our method and demonstrate the value of ghost kitchens\ncompared to conventional meal delivery systems. We show that both integrated\noptimization of cook scheduling and vehicle dispatching, as well as\nanticipation of future demand and decisions, are essential for successful\noperations. We further derive several managerial insights, amongst others, that\ncompanies should carefully consider the trade-off between fast delivery and\nfresh food.\n", "date": "2024-08-14", "categories": []} {"id": "2408.06993", "title": "LLMs can Schedule", "abstract": " The job shop scheduling problem (JSSP) remains a significant hurdle in\noptimizing production processes. This challenge involves efficiently allocating\njobs to a limited number of machines while minimizing factors like total\nprocessing time or job delays. While recent advancements in artificial\nintelligence have yielded promising solutions, such as reinforcement learning\nand graph neural networks, this paper explores the potential of Large Language\nModels (LLMs) for JSSP. We introduce the very first supervised 120k dataset\nspecifically designed to train LLMs for JSSP. Surprisingly, our findings\ndemonstrate that LLM-based scheduling can achieve performance comparable to\nother neural approaches. Furthermore, we propose a sampling method that\nenhances the effectiveness of LLMs in tackling JSSP.\n", "date": "2024-08-13", "categories": []} {"id": "2310.20609", "title": "Graph Matching via convex relaxation to the simplex", "abstract": " This paper addresses the Graph Matching problem, which consists of finding\nthe best possible alignment between two input graphs, and has many applications\nin computer vision, network deanonymization and protein alignment. A common\napproach to tackle this problem is through convex relaxations of the NP-hard\n\\emph{Quadratic Assignment Problem} (QAP).\n Here, we introduce a new convex relaxation onto the unit simplex and develop\nan efficient mirror descent scheme with closed-form iterations for solving this\nproblem. Under the correlated Gaussian Wigner model, we show that the simplex\nrelaxation admits a unique solution with high probability. In the noiseless\ncase, this is shown to imply exact recovery of the ground truth permutation.\nAdditionally, we establish a novel sufficiency condition for the input matrix\nin standard greedy rounding methods, which is less restrictive than the\ncommonly used `diagonal dominance' condition. We use this condition to show\nexact one-step recovery of the ground truth (holding almost surely) via the\nmirror descent scheme, in the noiseless setting. We also use this condition to\nobtain significantly improved conditions for the GRAMPA algorithm [Fan et al.\n2019] in the noiseless setting.\n", "date": "2023-10-31", "categories": []} {"id": "2408.03585", "title": "Hierarchical Neural Constructive Solver for Real-world TSP Scenarios", "abstract": " Existing neural constructive solvers for routing problems have predominantly\nemployed transformer architectures, conceptualizing the route construction as a\nset-to-sequence learning task. However, their efficacy has primarily been\ndemonstrated on entirely random problem instances that inadequately capture\nreal-world scenarios. In this paper, we introduce realistic Traveling Salesman\nProblem (TSP) scenarios relevant to industrial settings and derive the\nfollowing insights: (1) The optimal next node (or city) to visit often lies\nwithin proximity to the current node, suggesting the potential benefits of\nbiasing choices based on current locations. (2) Effectively solving the TSP\nrequires robust tracking of unvisited nodes and warrants succinct grouping\nstrategies. Building upon these insights, we propose integrating a learnable\nchoice layer inspired by Hypernetworks to prioritize choices based on the\ncurrent location, and a learnable approximate clustering algorithm inspired by\nthe Expectation-Maximization algorithm to facilitate grouping the unvisited\ncities. Together, these two contributions form a hierarchical approach towards\nsolving the realistic TSP by considering both immediate local neighbourhoods\nand learning an intermediate set of node representations. Our hierarchical\napproach yields superior performance compared to both classical and recent\ntransformer models, showcasing the efficacy of the key designs.\n", "date": "2024-08-07", "categories": []} {"id": "2408.03076", "title": "Solving QUBO on the Loihi 2 Neuromorphic Processor", "abstract": " In this article, we describe an algorithm for solving Quadratic Unconstrained\nBinary Optimization problems on the Intel Loihi 2 neuromorphic processor. The\nsolver is based on a hardware-aware fine-grained parallel simulated annealing\nalgorithm developed for Intel's neuromorphic research chip Loihi 2. Preliminary\nresults show that our approach can generate feasible solutions in as little as\n1 ms and up to 37x more energy efficient compared to two baseline solvers\nrunning on a CPU. These advantages could be especially relevant for size-,\nweight-, and power-constrained edge computing applications.\n", "date": "2024-08-06", "categories": []} {"id": "2408.02207", "title": "MARCO: A Memory-Augmented Reinforcement Framework for Combinatorial\n Optimization", "abstract": " Neural Combinatorial Optimization (NCO) is an emerging domain where deep\nlearning techniques are employed to address combinatorial optimization problems\nas a standalone solver. Despite their potential, existing NCO methods often\nsuffer from inefficient search space exploration, frequently leading to local\noptima entrapment or redundant exploration of previously visited states. This\npaper introduces a versatile framework, referred to as Memory-Augmented\nReinforcement for Combinatorial Optimization (MARCO), that can be used to\nenhance both constructive and improvement methods in NCO through an innovative\nmemory module. MARCO stores data collected throughout the optimization\ntrajectory and retrieves contextually relevant information at each state. This\nway, the search is guided by two competing criteria: making the best decision\nin terms of the quality of the solution and avoiding revisiting already\nexplored solutions. This approach promotes a more efficient use of the\navailable optimization budget. Moreover, thanks to the parallel nature of NCO\nmodels, several search threads can run simultaneously, all sharing the same\nmemory module, enabling an efficient collaborative exploration. Empirical\nevaluations, carried out on the maximum cut, maximum independent set and\ntravelling salesman problems, reveal that the memory module effectively\nincreases the exploration, enabling the model to discover diverse,\nhigher-quality solutions. MARCO achieves good performance in a low\ncomputational cost, establishing a promising new direction in the field of NCO.\n", "date": "2024-08-05", "categories": []} {"id": "2408.01297", "title": "Optimal Mixed Integer Linear Optimization Trained Multivariate\n Classification Trees", "abstract": " Multivariate decision trees are powerful machine learning tools for\nclassification and regression that attract many researchers and industry\nprofessionals. An optimal binary tree has two types of vertices, (i) branching\nvertices which have exactly two children and where datapoints are assessed on a\nset of discrete features and (ii) leaf vertices at which datapoints are given a\nprediction, and can be obtained by solving a biobjective optimization problem\nthat seeks to (i) maximize the number of correctly classified datapoints and\n(ii) minimize the number of branching vertices. Branching vertices are linear\ncombinations of training features and therefore can be thought of as\nhyperplanes. In this paper, we propose two cut-based mixed integer linear\noptimization (MILO) formulations for designing optimal binary classification\ntrees (leaf vertices assign discrete classes). Our models leverage on-the-fly\nidentification of minimal infeasible subsystems (MISs) from which we derive\ncutting planes that hold the form of packing constraints. We show theoretical\nimprovements on the strongest flow-based MILO formulation currently in the\nliterature and conduct experiments on publicly available datasets to show our\nmodels' ability to scale, strength against traditional branch and bound\napproaches, and robustness in out-of-sample test performance. Our code and data\nare available on GitHub.\n", "date": "2024-08-02", "categories": []} {"id": "2401.10467", "title": "Learning Backdoors for Mixed Integer Linear Programs with Contrastive\n Learning", "abstract": " Many real-world problems can be efficiently modeled as Mixed Integer Linear\nPrograms (MILPs) and solved with the Branch-and-Bound method. Prior work has\nshown the existence of MILP backdoors, small sets of variables such that\nprioritizing branching on them when possible leads to faster running times.\nHowever, finding high-quality backdoors that improve running times remains an\nopen question. Previous work learns to estimate the relative solver speed of\nrandomly sampled backdoors through ranking and then decide whether to use the\nhighest-ranked backdoor candidate. In this paper, we utilize the Monte-Carlo\ntree search method to collect backdoors for training, rather than relying on\nrandom sampling, and adapt a contrastive learning framework to train a Graph\nAttention Network model to predict backdoors. Our method, evaluated on several\ncommon MILP problem domains, demonstrates performance improvements over both\nGurobi and previous models.\n", "date": "2024-01-19", "categories": []} {"id": "2407.21729", "title": "ParLS-PBO: A Parallel Local Search Solver for Pseudo Boolean\n Optimization", "abstract": " As a broadly applied technique in numerous optimization problems, recently,\nlocal search has been employed to solve Pseudo-Boolean Optimization (PBO)\nproblem. A representative local search solver for PBO is LSPBO. In this paper,\nfirstly, we improve LSPBO by a dynamic scoring mechanism, which dynamically\nstrikes a balance between score on hard constraints and score on the objective\nfunction.\n Moreover, on top of this improved LSPBO , we develop the first parallel local\nsearch PBO solver. The main idea is to share good solutions among different\nthreads to guide the search, by maintaining a pool of feasible solutions. For\nevaluating solutions when updating the pool, we propose a function that\nconsiders both the solution quality and the diversity of the pool. Furthermore,\nwe calculate the polarity density in the pool to enhance the scoring function\nof local search. Our empirical experiments show clear benefits of the proposed\nparallel approach, making it competitive with the parallel version of the\nfamous commercial solver Gurobi.\n", "date": "2024-07-31", "categories": []} {"id": "2311.03583", "title": "Finding Increasingly Large Extremal Graphs with AlphaZero and Tabu\n Search", "abstract": " This work studies a central extremal graph theory problem inspired by a 1975\nconjecture of Erd\\H{o}s, which aims to find graphs with a given size (number of\nnodes) that maximize the number of edges without having 3- or 4-cycles. We\nformulate this problem as a sequential decision-making problem and compare\nAlphaZero, a neural network-guided tree search, with tabu search, a heuristic\nlocal search method. Using either method, by introducing a curriculum --\njump-starting the search for larger graphs using good graphs found at smaller\nsizes -- we improve the state-of-the-art lower bounds for several sizes. We\nalso propose a flexible graph-generation environment and a\npermutation-invariant network architecture for learning to search in the space\nof graphs.\n", "date": "2023-11-06", "categories": []} {"id": "2407.19736", "title": "Sensor Selection via GFlowNets: A Deep Generative Modeling Framework to\n Navigate Combinatorial Complexity", "abstract": " The performance of sensor arrays in sensing and wireless communications\nimproves with more elements, but this comes at the cost of increased energy\nconsumption and hardware expense. This work addresses the challenge of\nselecting $k$ sensor elements from a set of $m$ to optimize a generic\nQuality-of-Service metric. Evaluating all $\\binom{m}{k}$ possible sensor\nsubsets is impractical, leading to prior solutions using convex relaxations,\ngreedy algorithms, and supervised learning approaches. The current paper\nproposes a new framework that employs deep generative modeling, treating sensor\nselection as a deterministic Markov Decision Process where sensor subsets of\nsize $k$ arise as terminal states. Generative Flow Networks (GFlowNets) are\nemployed to model an action distribution conditioned on the state. Sampling\nactions from the aforementioned distribution ensures that the probability of\narriving at a terminal state is proportional to the performance of the\ncorresponding subset. Applied to a standard sensor selection scenario, the\ndeveloped approach outperforms popular methods which are based on convex\noptimization and greedy algorithms. Finally, a multiobjective formulation of\nthe proposed approach is adopted and applied on the sparse antenna array design\nfor Integrated Sensing and Communication (ISAC) systems. The multiobjective\nvariation is shown to perform well in managing the trade-off between radar and\ncommunication performance.\n", "date": "2024-07-29", "categories": []} {"id": "2407.18436", "title": "A Model for Combinatorial Dictionary Learning and Inference", "abstract": " We are often interested in decomposing complex, structured data into simple\ncomponents that explain the data. The linear version of this problem is\nwell-studied as dictionary learning and factor analysis. In this work, we\npropose a combinatorial model in which to study this question, motivated by the\nway objects occlude each other in a scene to form an image. First, we identify\na property we call \"well-structuredness\" of a set of low-dimensional components\nwhich ensures that no two components in the set are too similar. We show how\nwell-structuredness is sufficient for learning the set of latent components\ncomprising a set of sample instances. We then consider the problem: given a set\nof components and an instance generated from some unknown subset of them,\nidentify which parts of the instance arise from which components. We consider\ntwo variants: (1) determine the minimal number of components required to\nexplain the instance; (2) determine the correct explanation for as many\nlocations as possible. For the latter goal, we also devise a version that is\nrobust to adversarial corruptions, with just a slightly stronger assumption on\nthe components. Finally, we show that the learning problem is computationally\ninfeasible in the absence of any assumptions.\n", "date": "2024-07-26", "categories": []} {"id": "2407.16092", "title": "Faster Optimal Coalition Structure Generation via Offline Coalition\n Selection and Graph-Based Search", "abstract": " Coalition formation is a key capability in multi-agent systems. An important\nproblem in coalition formation is coalition structure generation: partitioning\nagents into coalitions to optimize the social welfare. This is a challenging\nproblem that has been the subject of active research for the past three\ndecades. In this paper, we present a novel algorithm, SMART, for the problem\nbased on a hybridization of three innovative techniques. Two of these\ntechniques are based on dynamic programming, where we show a powerful\nconnection between the coalitions selected for evaluation and the performance\nof the algorithms. These algorithms use offline phases to optimize the choice\nof coalitions to evaluate. The third one uses branch-and-bound and integer\npartition graph search to explore the solution space. Our techniques bring a\nnew way of approaching the problem and a new level of precision to the field.\nIn experiments over several common value distributions, we show that the\nhybridization of these techniques in SMART is faster than the fastest prior\nalgorithms (ODP-IP, BOSS) in generating optimal solutions across all the value\ndistributions.\n", "date": "2024-07-22", "categories": []} {"id": "2407.16200", "title": "MCTS Based Dispatch of Autonomous Vehicles under Operational Constraints\n for Continuous Transportation", "abstract": " Continuous transportation of material in the mining industry is achieved by\nthe dispatch of autonomous haul-trucks with discrete haulage capacities.\nRecently, Monte Carlo Tree Search (MCTS) was successfully deployed in tackling\nchallenges of long-run optimality, scalability and adaptability in haul-truck\ndispatch. Typically, operational constraints imposed on the mine site are\nsatisfied by heuristic controllers or human operators independent of the\ndispatch planning. This article incorporates operational constraint\nsatisfaction into the dispatch planning by utilising the MCTS based dispatch\nplanner Flow-Achieving Scheduling Tree (FAST). Operational constraint violation\nand satisfaction are modelled as opportunity costs in the combinatorial\noptimisation problem of dispatch. Explicit cost formulations are avoided by\nutilising MCTS generator models to derive opportunity costs. Experimental\nstudies with four types of operational constraints demonstrate the success of\nutilising opportunity costs for constraint satisfaction, and the effectiveness\nof integrating constraints into dispatch planning.\n", "date": "2024-07-23", "categories": []} {"id": "2407.16468", "title": "Enhancing GNNs Performance on Combinatorial Optimization by Recurrent\n Feature Update", "abstract": " Combinatorial optimization (CO) problems are crucial in various scientific\nand industrial applications. Recently, researchers have proposed using\nunsupervised Graph Neural Networks (GNNs) to address NP-hard combinatorial\noptimization problems, which can be reformulated as Quadratic Unconstrained\nBinary Optimization (QUBO) problems. GNNs have demonstrated high performance\nwith nearly linear scalability and significantly outperformed classic\nheuristic-based algorithms in terms of computational efficiency on large-scale\nproblems. However, when utilizing standard node features, GNNs tend to get\ntrapped to suboptimal local minima of the energy landscape, resulting in low\nquality solutions. We introduce a novel algorithm, denoted hereafter as\nQRF-GNN, leveraging the power of GNNs to efficiently solve CO problems with\nQUBO formulation. It relies on unsupervised learning by minimizing the loss\nfunction derived from QUBO relaxation. The proposed key components of the\narchitecture include the recurrent use of intermediate GNN predictions,\nparallel convolutional layers and combination of static node features as input.\nAltogether, it helps to adapt the intermediate solution candidate to minimize\nQUBO-based loss function, taking into account not only static graph features,\nbut also intermediate predictions treated as dynamic, i.e. iteratively changing\nrecurrent features. The performance of the proposed algorithm has been\nevaluated on the canonical benchmark datasets for maximum cut, graph coloring\nand maximum independent set problems. Results of experiments show that QRF-GNN\ndrastically surpasses existing learning-based approaches and is comparable to\nthe state-of-the-art conventional heuristics, improving their scalability on\nlarge instances.\n", "date": "2024-07-23", "categories": []} {"id": "2210.03918", "title": "Finding and Exploring Promising Search Space for the 0-1\n Multidimensional Knapsack Problem", "abstract": " The 0-1 Multidimensional Knapsack Problem (MKP) is a classical NP-hard\ncombinatorial optimization problem with many engineering applications. In this\npaper, we propose a novel algorithm combining evolutionary computation with the\nexact algorithm to solve the 0-1 MKP. It maintains a set of solutions and\nutilizes the information from the population to extract good partial\nassignments. To find high-quality solutions, an exact algorithm is applied to\nexplore the promising search space specified by the good partial assignments.\nThe new solutions are used to update the population. Thus, the good partial\nassignments evolve towards a better direction with the improvement of the\npopulation. Extensive experimentation with commonly used benchmark sets shows\nthat our algorithm outperforms the state-of-the-art heuristic algorithms, TPTEA\nand DQPSO, as well as the commercial solver CPlex. It finds better solutions\nthan the existing algorithms and provides new lower bounds for 10 large and\nhard instances.\n", "date": "2022-10-08", "categories": []} {"id": "2301.13395", "title": "Differentiating Through Integer Linear Programs with Quadratic\n Regularization and Davis-Yin Splitting", "abstract": " In many applications, a combinatorial problem must be repeatedly solved with\nsimilar, but distinct parameters. Yet, the parameters $w$ are not directly\nobserved; only contextual data $d$ that correlates with $w$ is available. It is\ntempting to use a neural network to predict $w$ given $d$. However, training\nsuch a model requires reconciling the discrete nature of combinatorial\noptimization with the gradient-based frameworks used to train neural networks.\nWe study the case where the problem in question is an Integer Linear Program\n(ILP). We propose applying a three-operator splitting technique, also known as\nDavis-Yin splitting (DYS), to the quadratically regularized continuous\nrelaxation of the ILP. We prove that the resulting scheme is compatible with\nthe recently introduced Jacobian-free backpropagation (JFB). Our experiments on\ntwo representative ILPs: the shortest path problem and the knapsack problem,\ndemonstrate that this combination-DYS on the forward pass, JFB on the backward\npass-yields a scheme which scales more effectively to high-dimensional problems\nthan existing schemes. All code associated with this paper is available at\ngithub.com/mines-opt-ml/fpo-dys.\n", "date": "2023-01-31", "categories": []} {"id": "2312.08224", "title": "GLOP: Learning Global Partition and Local Construction for Solving\n Large-scale Routing Problems in Real-time", "abstract": " The recent end-to-end neural solvers have shown promise for small-scale\nrouting problems but suffered from limited real-time scaling-up performance.\nThis paper proposes GLOP (Global and Local Optimization Policies), a unified\nhierarchical framework that efficiently scales toward large-scale routing\nproblems. GLOP partitions large routing problems into Travelling Salesman\nProblems (TSPs) and TSPs into Shortest Hamiltonian Path Problems. For the first\ntime, we hybridize non-autoregressive neural heuristics for coarse-grained\nproblem partitions and autoregressive neural heuristics for fine-grained route\nconstructions, leveraging the scalability of the former and the meticulousness\nof the latter. Experimental results show that GLOP achieves competitive and\nstate-of-the-art real-time performance on large-scale routing problems,\nincluding TSP, ATSP, CVRP, and PCTSP.\n", "date": "2023-12-13", "categories": []} {"id": "2406.14697", "title": "A Benchmark Study of Deep-RL Methods for Maximum Coverage Problems over\n Graphs", "abstract": " Recent years have witnessed a growing trend toward employing deep\nreinforcement learning (Deep-RL) to derive heuristics for combinatorial\noptimization (CO) problems on graphs. Maximum Coverage Problem (MCP) and its\nprobabilistic variant on social networks, Influence Maximization (IM), have\nbeen particularly prominent in this line of research. In this paper, we present\na comprehensive benchmark study that thoroughly investigates the effectiveness\nand efficiency of five recent Deep-RL methods for MCP and IM. These methods\nwere published in top data science venues, namely S2V-DQN, Geometric-QN, GCOMB,\nRL4IM, and LeNSE. Our findings reveal that, across various scenarios, the Lazy\nGreedy algorithm consistently outperforms all Deep-RL methods for MCP. In the\ncase of IM, theoretically sound algorithms like IMM and OPIM demonstrate\nsuperior performance compared to Deep-RL methods in most scenarios. Notably, we\nobserve an abnormal phenomenon in IM problem where Deep-RL methods slightly\noutperform IMM and OPIM when the influence spread nearly does not increase as\nthe budget increases. Furthermore, our experimental results highlight common\nissues when applying Deep-RL methods to MCP and IM in practical settings.\nFinally, we discuss potential avenues for improving Deep-RL methods. Our\nbenchmark study sheds light on potential challenges in current deep\nreinforcement learning research for solving combinatorial optimization\nproblems.\n", "date": "2024-06-20", "categories": []} {"id": "2407.13917", "title": "LinSATNet: The Positive Linear Satisfiability Neural Networks", "abstract": " Encoding constraints into neural networks is attractive. This paper studies\nhow to introduce the popular positive linear satisfiability to neural networks.\nWe propose the first differentiable satisfiability layer based on an extension\nof the classic Sinkhorn algorithm for jointly encoding multiple sets of\nmarginal distributions. We further theoretically characterize the convergence\nproperty of the Sinkhorn algorithm for multiple marginals. In contrast to the\nsequential decision e.g.\\ reinforcement learning-based solvers, we showcase our\ntechnique in solving constrained (specifically satisfiability) problems by\none-shot neural networks, including i) a neural routing solver learned without\nsupervision of optimal solutions; ii) a partial graph matching network handling\ngraphs with unmatchable outliers on both sides; iii) a predictive network for\nfinancial portfolios with continuous constraints. To our knowledge, there\nexists no one-shot neural solver for these scenarios when they are formulated\nas satisfiability problems. Source code is available at\nhttps://github.com/Thinklab-SJTU/LinSATNet\n", "date": "2024-07-18", "categories": []} {"id": "2208.11010", "title": "Convex mixed-integer optimization with Frank-Wolfe methods", "abstract": " Mixed-integer nonlinear optimization encompasses a broad class of problems\nthat present both theoretical and computational challenges. We propose a new\ntype of method to solve these problems based on a branch-and-bound algorithm\nwith convex node relaxations. These relaxations are solved with a Frank-Wolfe\nalgorithm over the convex hull of mixed-integer feasible points instead of the\ncontinuous relaxation via calls to a mixed-integer linear solver as the linear\nminimization oracle. The proposed method computes feasible solutions while\nworking on a single representation of the polyhedral constraints, leveraging\nthe full extent of mixed-integer linear solvers without an outer approximation\nscheme and can exploit inexact solutions of node subproblems.\n", "date": "2022-08-23", "categories": []} {"id": "2407.13023", "title": "A Three-Stage Algorithm for the Closest String Problem on Artificial and\n Real Gene Sequences", "abstract": " The Closest String Problem is an NP-hard problem that aims to find a string\nthat has the minimum distance from all sequences that belong to the given set\nof strings. Its applications can be found in coding theory, computational\nbiology, and designing degenerated primers, among others. There are efficient\nexact algorithms that have reached high-quality solutions for binary sequences.\nHowever, there is still room for improvement concerning the quality of\nsolutions over DNA and protein sequences. In this paper, we introduce a\nthree-stage algorithm that comprises the following process: first, we apply a\nnovel alphabet pruning method to reduce the search space for effectively\nfinding promising search regions. Second, a variant of beam search to find a\nheuristic solution is employed. This method utilizes a newly developed guiding\nfunction based on an expected distance heuristic score of partial solutions.\nLast, we introduce a local search to improve the quality of the solution\nobtained from the beam search. Furthermore, due to the lack of real-world\nbenchmarks, two real-world datasets are introduced to verify the robustness of\nthe method. The extensive experimental results show that the proposed method\noutperforms the previous approaches from the literature.\n", "date": "2024-07-17", "categories": []} {"id": "2407.13113", "title": "Multiobjective Vehicle Routing Optimization with Time Windows: A Hybrid\n Approach Using Deep Reinforcement Learning and NSGA-II", "abstract": " This paper proposes a weight-aware deep reinforcement learning (WADRL)\napproach designed to address the multiobjective vehicle routing problem with\ntime windows (MOVRPTW), aiming to use a single deep reinforcement learning\n(DRL) model to solve the entire multiobjective optimization problem. The\nNon-dominated sorting genetic algorithm-II (NSGA-II) method is then employed to\noptimize the outcomes produced by the WADRL, thereby mitigating the limitations\nof both approaches. Firstly, we design an MOVRPTW model to balance the\nminimization of travel cost and the maximization of customer satisfaction.\nSubsequently, we present a novel DRL framework that incorporates a\ntransformer-based policy network. This network is composed of an encoder\nmodule, a weight embedding module where the weights of the objective functions\nare incorporated, and a decoder module. NSGA-II is then utilized to optimize\nthe solutions generated by WADRL. Finally, extensive experimental results\ndemonstrate that our method outperforms the existing and traditional methods.\nDue to the numerous constraints in VRPTW, generating initial solutions of the\nNSGA-II algorithm can be time-consuming. However, using solutions generated by\nthe WADRL as initial solutions for NSGA-II significantly reduces the time\nrequired for generating initial solutions. Meanwhile, the NSGA-II algorithm can\nenhance the quality of solutions generated by WADRL, resulting in solutions\nwith better scalability. Notably, the weight-aware strategy significantly\nreduces the training time of DRL while achieving better results, enabling a\nsingle DRL model to solve the entire multiobjective optimization problem.\n", "date": "2024-07-18", "categories": []} {"id": "2407.11364", "title": "Learning-augmented Maximum Independent Set", "abstract": " We study the Maximum Independent Set (MIS) problem on general graphs within\nthe framework of learning-augmented algorithms. The MIS problem is known to be\nNP-hard and is also NP-hard to approximate to within a factor of $n^{1-\\delta}$\nfor any $\\delta>0$. We show that we can break this barrier in the presence of\nan oracle obtained through predictions from a machine learning model that\nanswers vertex membership queries for a fixed MIS with probability\n$1/2+\\varepsilon$. In the first setting we consider, the oracle can be queried\nonce per vertex to know if a vertex belongs to a fixed MIS, and the oracle\nreturns the correct answer with probability $1/2 + \\varepsilon$. Under this\nsetting, we show an algorithm that obtains an\n$\\tilde{O}(\\sqrt{\\Delta}/\\varepsilon)$-approximation in $O(m)$ time where\n$\\Delta$ is the maximum degree of the graph. In the second setting, we allow\nmultiple queries to the oracle for a vertex, each of which is correct with\nprobability $1/2 + \\varepsilon$. For this setting, we show an\n$O(1)$-approximation algorithm using $O(n/\\varepsilon^2)$ total queries and\n$\\tilde{O}(m)$ runtime.\n", "date": "2024-07-16", "categories": []} {"id": "2407.08633", "title": "A Novel Framework for Automated Warehouse Layout Generation", "abstract": " Optimizing warehouse layouts is crucial due to its significant impact on\nefficiency and productivity. We present an AI-driven framework for automated\nwarehouse layout generation. This framework employs constrained beam search to\nderive optimal layouts within given spatial parameters, adhering to all\nfunctional requirements. The feasibility of the generated layouts is verified\nbased on criteria such as item accessibility, required minimum clearances, and\naisle connectivity. A scoring function is then used to evaluate the feasible\nlayouts considering the number of storage locations, access points, and\naccessibility costs. We demonstrate our method's ability to produce feasible,\noptimal layouts for a variety of warehouse dimensions and shapes, diverse door\nplacements, and interconnections. This approach, currently being prepared for\ndeployment, will enable human designers to rapidly explore and confirm options,\nfacilitating the selection of the most appropriate layout for their use-case.\n", "date": "2024-07-11", "categories": []} {"id": "2407.10003", "title": "A Dynamic Algorithm for Weighted Submodular Cover Problem", "abstract": " We initiate the study of the submodular cover problem in dynamic setting\nwhere the elements of the ground set are inserted and deleted.\n In the classical submodular cover problem, we are given a monotone submodular\nfunction $f : 2^{V} \\to \\mathbb{R}^{\\ge 0}$ and the goal is to obtain a set $S\n\\subseteq V$ that minimizes the cost subject to the constraint $f(S) = f(V)$.\nThis is a classical problem in computer science and generalizes the Set Cover\nproblem, 2-Set Cover, and dominating set problem among others.\n We consider this problem in a dynamic setting where there are updates to our\nset $V$, in the form of insertions and deletions of elements from a ground set\n$\\mathcal{V}$, and the goal is to maintain an approximately optimal solution\nwith low query complexity per update. For this problem, we propose a randomized\nalgorithm that, in expectation, obtains a $(1-O(\\epsilon),\nO(\\epsilon^{-1}))$-bicriteria approximation using polylogarithmic query\ncomplexity per update.\n", "date": "2024-07-13", "categories": []} {"id": "2306.04647", "title": "Compressed Sensing: A Discrete Optimization Approach", "abstract": " We study the Compressed Sensing (CS) problem, which is the problem of finding\nthe most sparse vector that satisfies a set of linear measurements up to some\nnumerical tolerance. We introduce an $\\ell_2$ regularized formulation of CS\nwhich we reformulate as a mixed integer second order cone program. We derive a\nsecond order cone relaxation of this problem and show that under mild\nconditions on the regularization parameter, the resulting relaxation is\nequivalent to the well studied basis pursuit denoising problem. We present a\nsemidefinite relaxation that strengthens the second order cone relaxation and\ndevelop a custom branch-and-bound algorithm that leverages our second order\ncone relaxation to solve small-scale instances of CS to certifiable optimality.\nWhen compared against solutions produced by three state of the art benchmark\nmethods on synthetic data, our numerical results show that our approach\nproduces solutions that are on average $6.22\\%$ more sparse. When compared only\nagainst the experiment-wise best performing benchmark method on synthetic data,\nour approach produces solutions that are on average $3.10\\%$ more sparse. On\nreal world ECG data, for a given $\\ell_2$ reconstruction error our approach\nproduces solutions that are on average $9.95\\%$ more sparse than benchmark\nmethods ($3.88\\%$ more sparse if only compared against the best performing\nbenchmark), while for a given sparsity level our approach produces solutions\nthat have on average $10.77\\%$ lower reconstruction error than benchmark\nmethods ($1.42\\%$ lower error if only compared against the best performing\nbenchmark). When used as a component of a multi-label classification algorithm,\nour approach achieves greater classification accuracy than benchmark compressed\nsensing methods. This improved accuracy comes at the cost of an increase in\ncomputation time by several orders of magnitude.\n", "date": "2023-06-05", "categories": []} {"id": "2407.07355", "title": "High-Precision, Fair University Course Scheduling During a Pandemic", "abstract": " Scheduling university courses is extra challenging when classroom capacities\nare reduced because of social distancing requirements that are implemented in\nresponse to a pandemic such as COVID-19. In this work, we propose an expanded\ntaxonomy of course delivery modes, present an integer program, and develop a\ncourse scheduling algorithm to enable all course sections -- even the largest\n-- to have a significant classroom learning component during a pandemic. Our\napproach is fair by ensuring that a certain fraction of the instruction in\nevery course section occurs in the classroom. Unlike previous studies, we do\nnot allow rotating attendance and instead require simultaneous attendance in\nwhich all students in a section meet in 1-5 rooms at the same time but less\noften than in a normal semester. These mass meetings, which create\nopportunities for in-person midterm exams and group activities, are scheduled\nat high precision across all days of the semester rather than a single,\nrepeating week. A fast heuristic algorithm makes the schedule in an hour.\nResults: We consider the 1834 in-person course sections, 172 classrooms, and 96\ndays in the fall 2022 semester at [UniversityXYZ]. If average classroom\ncapacity is reduced by 75% due to a pandemic, our approach still allows at\nleast 25% of the instruction in every section, and more than 49% of all\ninstruction across the entire campus, to be in the classroom. Our method also\nproduces excellent results for regular classroom assignment. Managerial\nimplications: An algorithm based on the principles of fairness and simultaneous\nattendance can significantly improve university course schedules during a\npandemic and in normal times. High-precision schedules that prepare a campus\nfor various pandemic possibilities can be created with minimal administrative\neffort and activated at a moment's notice before or during a semester if an\noutbreak occurs.\n", "date": "2024-07-10", "categories": []} {"id": "2406.09740", "title": "Deep Symbolic Optimization for Combinatorial Optimization: Accelerating\n Node Selection by Discovering Potential Heuristics", "abstract": " Combinatorial optimization (CO) is one of the most fundamental mathematical\nmodels in real-world applications. Traditional CO solvers, such as\nBranch-and-Bound (B&B) solvers, heavily rely on expert-designed heuristics,\nwhich are reliable but require substantial manual tuning. Recent studies have\nleveraged deep learning (DL) models as an alternative to capture rich feature\npatterns for improved performance on GPU machines. Nonetheless, the drawbacks\nof high training and inference costs, as well as limited interpretability,\nseverely hinder the adoption of DL methods in real-world applications. To\naddress these challenges, we propose a novel deep symbolic optimization\nlearning framework that combines their advantages. Specifically, we focus on\nthe node selection module within B&B solvers -- namely, deep symbolic\noptimization for node selection (Dso4NS). With data-driven approaches, Dso4NS\nguides the search for mathematical expressions within the high-dimensional\ndiscrete symbolic space and then incorporates the highest-performing\nmathematical expressions into a solver. The data-driven model captures the rich\nfeature information in the input data and generates symbolic expressions, while\nthe expressions deployed in solvers enable fast inference with high\ninterpretability. Experiments demonstrate the effectiveness of Dso4NS in\nlearning high-quality expressions, outperforming existing approaches on a CPU\nmachine. Encouragingly, the learned CPU-based policies consistently achieve\nperformance comparable to state-of-the-art GPU-based approaches.\n", "date": "2024-06-14", "categories": []} {"id": "2310.04471", "title": "Fast Neighborhood Search Heuristics for the Colored Bin Packing Problem", "abstract": " The Colored Bin Packing Problem (CBPP) is a generalization of the Bin Packing\nProblem (BPP). The CBPP consists of packing a set of items, each with a weight\nand a color, in bins of limited capacity, minimizing the number of used bins\nand satisfying the constraint that two items of the same color cannot be packed\nside by side in the same bin. In this article, we proposed an adaptation of BPP\nheuristics and new heuristics for the CBPP. Moreover, we propose a set of fast\nneighborhood search algorithms for CBPP. These neighborhoods are applied in a\nmeta-heuristic approach based on the Variable Neighborhood Search (VNS) and a\nmatheuristic approach that combines linear programming with the meta-heuristics\nVNS and Greedy Randomized Adaptive Search (GRASP). The results indicate that\nour matheuristic is superior to VNS and that both approaches can find\nnear-optimal solutions for a large number of instances, even for those with\nmany items.\n", "date": "2023-10-06", "categories": []} {"id": "2308.11038", "title": "Logistics Hub Location Optimization: A K-Means and P-Median Model Hybrid\n Approach Using Road Network Distances", "abstract": " Logistic hubs play a pivotal role in the last-mile delivery distance; even a\nslight increment in distance negatively impacts the business of the e-commerce\nindustry while also increasing its carbon footprint. The growth of this\nindustry, particularly after Covid-19, has further intensified the need for\noptimized allocation of resources in an urban environment. In this study, we\nuse a hybrid approach to optimize the placement of logistic hubs. The approach\nsequentially employs different techniques. Initially, delivery points are\nclustered using K-Means in relation to their spatial locations. The clustering\nmethod utilizes road network distances as opposed to Euclidean distances.\nNon-road network-based approaches have been avoided since they lead to\nerroneous and misleading results. Finally, hubs are located using the P-Median\nmethod. The P-Median method also incorporates the number of deliveries and\npopulation as weights. Real-world delivery data from Muller and Phipps (M&P) is\nused to demonstrate the effectiveness of the approach. Serving deliveries from\nthe optimal hub locations results in the saving of 815 (10%) meters per\ndelivery.\n", "date": "2023-08-18", "categories": []} {"id": "2407.03927", "title": "Dancing to the State of the Art? How Candidate Lists Influence LKH for\n Solving the Traveling Salesperson Problem", "abstract": " Solving the Traveling Salesperson Problem (TSP) remains a persistent\nchallenge, despite its fundamental role in numerous generalized applications in\nmodern contexts. Heuristic solvers address the demand for finding high-quality\nsolutions efficiently. Among these solvers, the Lin-Kernighan-Helsgaun (LKH)\nheuristic stands out, as it complements the performance of genetic algorithms\nacross a diverse range of problem instances. However, frequent timeouts on\nchallenging instances hinder the practical applicability of the solver.\n Within this work, we investigate a previously overlooked factor contributing\nto many timeouts: The use of a fixed candidate set based on a tree structure.\nOur investigations reveal that candidate sets based on Hamiltonian circuits\ncontain more optimal edges. We thus propose to integrate this promising\ninitialization strategy, in the form of POPMUSIC, within an efficient restart\nversion of LKH. As confirmed by our experimental studies, this refined TSP\nheuristic is much more efficient - causing fewer timeouts and improving the\nperformance (in terms of penalized average runtime) by an order of magnitude -\nand thereby challenges the state of the art in TSP solving.\n", "date": "2024-07-04", "categories": []} {"id": "2302.06582", "title": "A Convex Hull Cheapest Insertion Heuristic for the Non-Euclidean TSP", "abstract": " The convex hull cheapest insertion heuristic produces good solutions to the\nEuclidean Traveling Salesperson Problem, but it has never been extended to the\nnon-Euclidean problem. This paper uses multidimensional scaling to first\nproject the points from a non-Euclidean space into a Euclidean space, enabling\nthe generation of a convex hull that initializes the algorithm. To evaluate the\nproposed algorithm, non-Euclidean spaces are created by adding separators to\nthe TSPLIB data-set, or by using the L1 norm as a metric.\n", "date": "2023-02-05", "categories": []} {"id": "2407.02601", "title": "Linear Submodular Maximization with Bandit Feedback", "abstract": " Submodular optimization with bandit feedback has recently been studied in a\nvariety of contexts. In a number of real-world applications such as diversified\nrecommender systems and data summarization, the submodular function exhibits\nadditional linear structure. We consider developing approximation algorithms\nfor the maximization of a submodular objective function\n$f:2^U\\to\\mathbb{R}_{\\geq 0}$, where $f=\\sum_{i=1}^dw_iF_{i}$. It is assumed\nthat we have value oracle access to the functions $F_i$, but the coefficients\n$w_i$ are unknown, and $f$ can only be accessed via noisy queries. We develop\nalgorithms for this setting inspired by adaptive allocation algorithms in the\nbest-arm identification for linear bandit, with approximation guarantees\narbitrarily close to the setting where we have value oracle access to $f$.\nFinally, we empirically demonstrate that our algorithms make vast improvements\nin terms of sample efficiency compared to algorithms that do not exploit the\nlinear structure of $f$ on instances of move recommendation.\n", "date": "2024-07-02", "categories": []} {"id": "2407.03180", "title": "A multi-objective combinatorial optimisation framework for large scale\n hierarchical population synthesis", "abstract": " In agent-based simulations, synthetic populations of agents are commonly used\nto represent the structure, behaviour, and interactions of individuals.\nHowever, generating a synthetic population that accurately reflects real\npopulation statistics is a challenging task, particularly when performed at\nscale. In this paper, we propose a multi objective combinatorial optimisation\ntechnique for large scale population synthesis. We demonstrate the\neffectiveness of our approach by generating a synthetic population for selected\nregions and validating it on contingency tables from real population data. Our\napproach supports complex hierarchical structures between individuals and\nhouseholds, is scalable to large populations and achieves minimal contigency\ntable reconstruction error. Hence, it provides a useful tool for policymakers\nand researchers for simulating the dynamics of complex populations.\n", "date": "2024-07-03", "categories": []} {"id": "2407.01546", "title": "Machine Learning-Enhanced Ant Colony Optimization for Column Generation", "abstract": " Column generation (CG) is a powerful technique for solving optimization\nproblems that involve a large number of variables or columns. This technique\nbegins by solving a smaller problem with a subset of columns and gradually\ngenerates additional columns as needed. However, the generation of columns\noften requires solving difficult subproblems repeatedly, which can be a\nbottleneck for CG. To address this challenge, we propose a novel method called\nmachine learning enhanced ant colony optimization (MLACO), to efficiently\ngenerate multiple high-quality columns from a subproblem. Specifically, we\ntrain a ML model to predict the optimal solution of a subproblem, and then\nintegrate this ML prediction into the probabilistic model of ACO to sample\nmultiple high-quality columns. Our experimental results on the bin packing\nproblem with conflicts show that the MLACO method significantly improves the\nperformance of CG compared to several state-of-the-art methods. Furthermore,\nwhen our method is incorporated into a Branch-and-Price method, it leads to a\nsignificant reduction in solution time.\n", "date": "2024-04-23", "categories": []} {"id": "2407.01615", "title": "Edge-DIRECT: A Deep Reinforcement Learning-based Method for Solving\n Heterogeneous Electric Vehicle Routing Problem with Time Window Constraints", "abstract": " In response to carbon-neutral policies in developed countries, electric\nvehicles route optimization has gained importance for logistics companies. With\nthe increasing focus on customer expectations and the shift towards more\ncustomer-oriented business models, the integration of delivery time-windows has\nbecome essential in logistics operations. Recognizing the critical nature of\nthese developments, this article studies the heterogeneous electric vehicle\nrouting problem with time-window constraints (HEVRPTW). To solve this variant\nof vehicle routing problem (VRP), we propose a DRL-based approach, named\nEdge-enhanced Dual attentIon encoderR and feature-EnhanCed dual aTtention\ndecoder (Edge-DIRECT). Edge-DIRECT features an extra graph representation, the\nnode connectivity of which is based on the overlap of customer time-windows.\nEdge-DIRECT's self-attention encoding mechanism is enhanced by exploiting the\nenergy consumption and travel time between the locations. To effectively\naccount for the heterogeneity of the EVs' fleet, a dual attention decoder has\nbeen introduced. Experimental results based on two real-world datasets reveal\nthat Edge-DIRECT outperforms a state-of-the-art DRL-based method and a\nwell-established heuristic approach in solution quality and execution time.\nFurthermore, it exhibits competitive performance when compared to another\nleading heuristic method.\n", "date": "2024-06-28", "categories": []} {"id": "2407.02091", "title": "Efficient Bit Labeling in Factorization Machines with Annealing for\n Traveling Salesman Problem", "abstract": " To efficiently find an optimum parameter combination in a large-scale\nproblem, it is a key to convert the parameters into available variables in\nactual machines. Specifically, quadratic unconstrained binary optimization\nproblems are solved with the help of machine learning, e.g., factorization\nmachines with annealing, which convert a raw parameter to binary variables.\nThis work investigates the dependence of the convergence speed and the accuracy\non binary labeling method, which can influence the cost function shape and thus\nthe probability of being captured at a local minimum solution. By exemplifying\ntraveling salesman problem, we propose and evaluate Gray labeling, which\ncorrelates the Hamming distance in binary labels with the traveling distance.\nThrough numerical simulation of traveling salesman problem up to 15 cities at a\nlimited number of iterations, the Gray labeling shows less local minima\npercentages and shorter traveling distances compared with natural labeling.\n", "date": "2024-07-02", "categories": []} {"id": "2407.00032", "title": "Design a Win-Win Strategy That Is Fair to Both Service Providers and\n Tasks When Rejection Is Not an Option", "abstract": " Assigning tasks to service providers is a frequent procedure across various\napplications. Often the tasks arrive dynamically while the service providers\nremain static. Preventing task rejection caused by service provider overload is\nof utmost significance. To ensure a positive experience in relevant\napplications for both service providers and tasks, fairness must be considered.\nTo address the issue, we model the problem as an online matching within a\nbipartite graph and tackle two minimax problems: one focuses on minimizing the\nhighest waiting time of a task, while the other aims to minimize the highest\nworkload of a service provider. We show that the second problem can be\nexpressed as a linear program and thus solved efficiently while maintaining a\nreasonable approximation to the objective of the first problem. We developed\nnovel methods that utilize the two minimax problems. We conducted extensive\nsimulation experiments using real data and demonstrated that our novel\nheuristics, based on the linear program, performed remarkably well.\n", "date": "2024-05-22", "categories": []} {"id": "2407.01440", "title": "GAT-Steiner: Rectilinear Steiner Minimal Tree Prediction Using GNNs", "abstract": " The Rectilinear Steiner Minimum Tree (RSMT) problem is a fundamental problem\nin VLSI placement and routing and is known to be NP-hard. Traditional RSMT\nalgorithms spend a significant amount of time on finding Steiner points to\nreduce the total wire length or use heuristics to approximate producing\nsub-optimal results. We show that Graph Neural Networks (GNNs) can be used to\npredict optimal Steiner points in RSMTs with high accuracy and can be\nparallelized on GPUs. In this paper, we propose GAT-Steiner, a graph attention\nnetwork model that correctly predicts 99.846% of the nets in the ISPD19\nbenchmark with an average increase in wire length of only 0.480% on suboptimal\nwire length nets. On randomly generated benchmarks, GAT-Steiner correctly\npredicts 99.942% with an average increase in wire length of only 0.420% on\nsuboptimal wire length nets.\n", "date": "2024-07-01", "categories": []} {"id": "2403.15114", "title": "Solving a Real-World Package Delivery Routing Problem Using Quantum\n Annealers", "abstract": " Research focused on the conjunction between quantum computing and routing\nproblems has been very prolific in recent years. Most of the works revolve\naround classical problems such as the Traveling Salesman Problem or the Vehicle\nRouting Problem. The real-world applicability of these problems is dependent on\nthe objectives and constraints considered. Anyway, it is undeniable that it is\noften difficult to translate complex requirements into these classical\nformulations.The main objective of this research is to present a solving scheme\nfor dealing with realistic instances while maintaining all the characteristics\nand restrictions of the original real-world problem. Thus, a quantum-classical\nstrategy has been developed, coined Q4RPD, that considers a set of real\nconstraints such as a heterogeneous fleet of vehicles, priority deliveries, and\ncapacities characterized by two values: weight and dimensions of the packages.\nQ4RPD resorts to the Leap Constrained Quadratic Model Hybrid Solver of D-Wave.\nTo demonstrate the application of Q4RPD, an experimentation composed of six\ndifferent instances has been conducted, aiming to serve as illustrative\nexamples.\n", "date": "2024-03-22", "categories": []} {"id": "2406.18781", "title": "Learning to Remove Cuts in Integer Linear Programming", "abstract": " Cutting plane methods are a fundamental approach for solving integer linear\nprograms (ILPs). In each iteration of such methods, additional linear\nconstraints (cuts) are introduced to the constraint set with the aim of\nexcluding the previous fractional optimal solution while not affecting the\noptimal integer solution. In this work, we explore a novel approach within\ncutting plane methods: instead of only adding new cuts, we also consider the\nremoval of previous cuts introduced at any of the preceding iterations of the\nmethod under a learnable parametric criteria. We demonstrate that in\nfundamental combinatorial optimization settings such cut removal policies can\nlead to significant improvements over both human-based and machine\nlearning-guided cut addition policies even when implemented with simple models.\n", "date": "2024-06-26", "categories": []} {"id": "2406.18119", "title": "Robust personnel rostering: how accurate should absenteeism predictions\n be?", "abstract": " Disruptions to personnel rosters caused by absenteeism often necessitate\nlast-minute adjustments to the employees' working hours. A common strategy to\nmitigate the impact of such changes is to assign employees to reserve shifts:\nspecial on-call duties during which an employee can be called in to cover for\nan absent employee. To maximize roster robustness, we assume a\npredict-then-optimize approach that uses absence predictions from a machine\nlearning model to schedule an adequate number of reserve shifts. In this paper\nwe propose a methodology to evaluate the robustness of rosters generated by the\npredict-then-optimize approach, assuming the machine learning model will make\npredictions at a predetermined prediction performance level. Instead of\ntraining and testing machine learning models, our methodology simulates the\npredictions based on a characterization of model performance. We show how this\nmethodology can be applied to identify the minimum performance level needed for\nthe model to outperform simple non-data-driven robust rostering policies. In a\ncomputational study on a nurse rostering problem, we demonstrate how the\npredict-then-optimize approach outperforms non-data-driven policies under\nreasonable performance requirements, particularly when employees possess\ninterchangeable skills.\n", "date": "2024-06-26", "categories": []} {"id": "2406.17334", "title": "Joint Admission Control and Resource Allocation of Virtual Network\n Embedding via Hierarchical Deep Reinforcement Learning", "abstract": " As an essential resource management problem in network virtualization,\nvirtual network embedding (VNE) aims to allocate the finite resources of\nphysical network to sequentially arriving virtual network requests (VNRs) with\ndifferent resource demands. Since this is an NP-hard combinatorial optimization\nproblem, many efforts have been made to provide viable solutions. However, most\nexisting approaches have either ignored the admission control of VNRs, which\nhas a potential impact on long-term performances, or not fully exploited the\ntemporal and topological features of the physical network and VNRs. In this\npaper, we propose a deep Hierarchical Reinforcement Learning approach to learn\na joint Admission Control and Resource Allocation policy for VNE, named\nHRL-ACRA. Specifically, the whole VNE process is decomposed into an upper-level\npolicy for deciding whether to admit the arriving VNR or not and a lower-level\npolicy for allocating resources of the physical network to meet the requirement\nof VNR through the HRL approach. Considering the proximal policy optimization\nas the basic training algorithm, we also adopt the average reward method to\naddress the infinite horizon problem of the upper-level agent and design a\ncustomized multi-objective intrinsic reward to alleviate the sparse reward\nissue of the lower-level agent. Moreover, we develop a deep feature-aware graph\nneural network to capture the features of VNR and physical network and exploit\na sequence-to-sequence model to generate embedding actions iteratively.\nFinally, extensive experiments are conducted in various settings, and show that\nHRL-ACRA outperforms state-of-the-art baselines in terms of both the acceptance\nratio and long-term average revenue. Our code is available at\n\\url{https://github.com/GeminiLight/hrl-acra}.\n", "date": "2024-06-25", "categories": []} {"id": "2406.14635", "title": "Harvesting Efficient On-Demand Order Pooling from Skilled Couriers:\n Enhancing Graph Representation Learning for Refining Real-time Many-to-One\n Assignments", "abstract": " The recent past has witnessed a notable surge in on-demand food delivery\n(OFD) services, offering delivery fulfillment within dozens of minutes after an\norder is placed. In OFD, pooling multiple orders for simultaneous delivery in\nreal-time order assignment is a pivotal efficiency source, which may in turn\nextend delivery time. Constructing high-quality order pooling to harmonize\nplatform efficiency with the experiences of consumers and couriers, is crucial\nto OFD platforms. However, the complexity and real-time nature of order\nassignment, making extensive calculations impractical, significantly limit the\npotential for order consolidation. Moreover, offline environment is frequently\nriddled with unknown factors, posing challenges for the platform's\nperceptibility and pooling decisions. Nevertheless, delivery behaviors of\nskilled couriers (SCs) who know the environment well, can improve system\nawareness and effectively inform decisions. Hence a SC delivery network (SCDN)\nis constructed, based on an enhanced attributed heterogeneous network embedding\napproach tailored for OFD. It aims to extract features from rich temporal and\nspatial information, and uncover the latent potential for order combinations\nembedded within SC trajectories. Accordingly, the vast search space of order\nassignment can be effectively pruned through scalable similarity calculations\nof low-dimensional vectors, making comprehensive and high-quality pooling\noutcomes more easily identified in real time. SCDN has now been deployed in\nMeituan dispatch system. Online tests reveal that with SCDN, the pooling\nquality and extent have been greatly improved. And our system can boost\ncouriers'efficiency by 45-55% during noon peak hours, while upholding the\ntimely delivery commitment.\n", "date": "2024-06-20", "categories": []} {"id": "2307.07357", "title": "Inverse Optimization for Routing Problems", "abstract": " We propose a method for learning decision-makers' behavior in routing\nproblems using Inverse Optimization (IO). The IO framework falls into the\nsupervised learning category and builds on the premise that the target behavior\nis an optimizer of an unknown cost function. This cost function is to be\nlearned through historical data, and in the context of routing problems, can be\ninterpreted as the routing preferences of the decision-makers. In this view,\nthe main contributions of this study are to propose an IO methodology with a\nhypothesis function, loss function, and stochastic first-order algorithm\ntailored to routing problems. We further test our IO approach in the Amazon\nLast Mile Routing Research Challenge, where the goal is to learn models that\nreplicate the routing preferences of human drivers, using thousands of\nreal-world routing examples. Our final IO-learned routing model achieves a\nscore that ranks 2nd compared with the 48 models that qualified for the final\nround of the challenge. Our examples and results showcase the flexibility and\nreal-world potential of the proposed IO methodology to learn from\ndecision-makers' decisions in routing problems.\n", "date": "2023-07-14", "categories": []} {"id": "2406.09899", "title": "Learning Solution-Aware Transformers for Efficiently Solving Quadratic\n Assignment Problem", "abstract": " Recently various optimization problems, such as Mixed Integer Linear\nProgramming Problems (MILPs), have undergone comprehensive investigation,\nleveraging the capabilities of machine learning. This work focuses on\nlearning-based solutions for efficiently solving the Quadratic Assignment\nProblem (QAPs), which stands as a formidable challenge in combinatorial\noptimization. While many instances of simpler problems admit fully\npolynomial-time approximate solution (FPTAS), QAP is shown to be strongly\nNP-hard. Even finding a FPTAS for QAP is difficult, in the sense that the\nexistence of a FPTAS implies $P = NP$. Current research on QAPs suffer from\nlimited scale and computational inefficiency. To attack the aforementioned\nissues, we here propose the first solution of its kind for QAP in the\nlearn-to-improve category. This work encodes facility and location nodes\nseparately, instead of forming computationally intensive association graphs\nprevalent in current approaches. This design choice enables scalability to\nlarger problem sizes. Furthermore, a \\textbf{S}olution \\textbf{AW}are\n\\textbf{T}ransformer (SAWT) architecture integrates the incumbent solution\nmatrix with the attention score to effectively capture higher-order information\nof the QAPs. Our model's effectiveness is validated through extensive\nexperiments on self-generated QAP instances of varying sizes and the QAPLIB\nbenchmark.\n", "date": "2024-06-14", "categories": []} {"id": "2406.13125", "title": "A Unified Framework for Combinatorial Optimization Based on Graph Neural\n Networks", "abstract": " Graph neural networks (GNNs) have emerged as a powerful tool for solving\ncombinatorial optimization problems (COPs), exhibiting state-of-the-art\nperformance in both graph-structured and non-graph-structured domains. However,\nexisting approaches lack a unified framework capable of addressing a wide range\nof COPs. After presenting a summary of representative COPs and a brief review\nof recent advancements in GNNs for solving COPs, this paper proposes a unified\nframework for solving COPs based on GNNs, including graph representation of\nCOPs, equivalent conversion of non-graph structured COPs to graph-structured\nCOPs, graph decomposition, and graph simplification. The proposed framework\nleverages the ability of GNNs to effectively capture the relational information\nand extract features from the graph representation of COPs, offering a generic\nsolution to COPs that can address the limitations of state-of-the-art in\nsolving non-graph-structured and highly complex graph-structured COPs.\n", "date": "2024-06-19", "categories": []} {"id": "2406.13414", "title": "Archive-based Single-Objective Evolutionary Algorithms for Submodular\n Optimization", "abstract": " Constrained submodular optimization problems play a key role in the area of\ncombinatorial optimization as they capture many NP-hard optimization problems.\nSo far, Pareto optimization approaches using multi-objective formulations have\nbeen shown to be successful to tackle these problems while single-objective\nformulations lead to difficulties for algorithms such as the $(1+1)$-EA due to\nthe presence of local optima. We introduce for the first time single-objective\nalgorithms that are provably successful for different classes of constrained\nsubmodular maximization problems. Our algorithms are variants of the\n$(1+\\lambda)$-EA and $(1+1)$-EA and increase the feasible region of the search\nspace incrementally in order to deal with the considered submodular problems.\n", "date": "2024-06-19", "categories": []} {"id": "2404.11677", "title": "Cross-Problem Learning for Solving Vehicle Routing Problems", "abstract": " Existing neural heuristics often train a deep architecture from scratch for\neach specific vehicle routing problem (VRP), ignoring the transferable\nknowledge across different VRP variants. This paper proposes the cross-problem\nlearning to assist heuristics training for different downstream VRP variants.\nParticularly, we modularize neural architectures for complex VRPs into 1) the\nbackbone Transformer for tackling the travelling salesman problem (TSP), and 2)\nthe additional lightweight modules for processing problem-specific features in\ncomplex VRPs. Accordingly, we propose to pre-train the backbone Transformer for\nTSP, and then apply it in the process of fine-tuning the Transformer models for\neach target VRP variant. On the one hand, we fully fine-tune the trained\nbackbone Transformer and problem-specific modules simultaneously. On the other\nhand, we only fine-tune small adapter networks along with the modules, keeping\nthe backbone Transformer still. Extensive experiments on typical VRPs\nsubstantiate that 1) the full fine-tuning achieves significantly better\nperformance than the one trained from scratch, and 2) the adapter-based\nfine-tuning also delivers comparable performance while being notably\nparameter-efficient. Furthermore, we empirically demonstrate the favorable\neffect of our method in terms of cross-distribution application and\nversatility.\n", "date": "2024-04-17", "categories": []} {"id": "2406.11897", "title": "A Benchmark for Maximum Cut: Towards Standardization of the Evaluation\n of Learned Heuristics for Combinatorial Optimization", "abstract": " Recently, there has been much work on the design of general heuristics for\ngraph-based, combinatorial optimization problems via the incorporation of Graph\nNeural Networks (GNNs) to learn distribution-specific solution\nstructures.However, there is a lack of consistency in the evaluation of these\nheuristics, in terms of the baselines and instances chosen, which makes it\ndifficult to assess the relative performance of the algorithms. In this paper,\nwe propose an open-source benchmark suite MaxCut-Bench dedicated to the NP-hard\nMaximum Cut problem in both its weighted and unweighted variants, based on a\ncareful selection of instances curated from diverse graph datasets. The suite\noffers a unified interface to various heuristics, both traditional and machine\nlearning-based. Next, we use the benchmark in an attempt to systematically\ncorroborate or reproduce the results of several, popular learning-based\napproaches, including S2V-DQN [31], ECO-DQN [4], among others, in terms of\nthree dimensions: objective value, generalization, and scalability. Our\nempirical results show that several of the learned heuristics fail to\noutperform a naive greedy algorithm, and that only one of them consistently\noutperforms Tabu Search, a simple, general heuristic based upon local search.\nFurthermore, we find that the performance of ECO-DQN remains the same or is\nimproved if the GNN is replaced by a simple linear regression on a subset of\nthe features that are related to Tabu Search. Code, data, and pretrained models\nare available at: \\url{https://github.com/ankurnath/MaxCut-Bench}.\n", "date": "2024-06-14", "categories": []} {"id": "2406.12349", "title": "Effective Generation of Feasible Solutions for Integer Programming via\n Guided Diffusion", "abstract": " Feasible solutions are crucial for Integer Programming (IP) since they can\nsubstantially speed up the solving process. In many applications, similar IP\ninstances often exhibit similar structures and shared solution distributions,\nwhich can be potentially modeled by deep learning methods. Unfortunately,\nexisting deep-learning-based algorithms, such as Neural Diving and\nPredict-and-search framework, are limited to generating only partial feasible\nsolutions, and they must rely on solvers like SCIP and Gurobi to complete the\nsolutions for a given IP problem. In this paper, we propose a novel framework\nthat generates complete feasible solutions end-to-end. Our framework leverages\ncontrastive learning to characterize the relationship between IP instances and\nsolutions, and learns latent embeddings for both IP instances and their\nsolutions. Further, the framework employs diffusion models to learn the\ndistribution of solution embeddings conditioned on IP representations, with a\ndedicated guided sampling strategy that accounts for both constraints and\nobjectives. We empirically evaluate our framework on four typical datasets of\nIP problems, and show that it effectively generates complete feasible solutions\nwith a high probability (> 89.7 \\%) without the reliance of Solvers and the\nquality of solutions is comparable to the best heuristic solutions from Gurobi.\nFurthermore, by integrating our method's sampled partial solutions with the\nCompleteSol heuristic from SCIP, the resulting feasible solutions outperform\nthose from state-of-the-art methods across all datasets, exhibiting a 3.7 to\n33.7\\% improvement in the gap to optimal values, and maintaining a feasible\nratio of over 99.7\\% for all datasets.\n", "date": "2024-06-18", "categories": []} {"id": "2406.12454", "title": "A Neural Column Generation Approach to the Vehicle Routing Problem with\n Two-Dimensional Loading and Last-In-First-Out Constraints", "abstract": " The vehicle routing problem with two-dimensional loading constraints\n(2L-CVRP) and the last-in-first-out (LIFO) rule presents significant practical\nand algorithmic challenges. While numerous heuristic approaches have been\nproposed to address its complexity, stemming from two NP-hard problems: the\nvehicle routing problem (VRP) and the two-dimensional bin packing problem\n(2D-BPP), less attention has been paid to developing exact algorithms. Bridging\nthis gap, this article presents an exact algorithm that integrates advanced\nmachine learning techniques, specifically a novel combination of attention and\nrecurrence mechanisms. This integration accelerates the state-of-the-art exact\nalgorithm by a median of 29.79% across various problem instances. Moreover, the\nproposed algorithm successfully resolves an open instance in the standard\ntest-bed, demonstrating significant improvements brought about by the\nincorporation of machine learning models. Code is available at\nhttps://github.com/xyfffff/NCG-for-2L-CVRP.\n", "date": "2024-06-18", "categories": []} {"id": "2406.09441", "title": "Comment on paper: Position: Rethinking Post-Hoc Search-Based Neural\n Approaches for Solving Large-Scale Traveling Salesman Problems", "abstract": " We identify two major issues in the SoftDist paper (Xia et al.): (1) the\nfailure to run all steps of different baselines on the same hardware\nenvironment, and (2) the use of inconsistent time measurements when comparing\nto other baselines. These issues lead to flawed conclusions. When all steps are\nexecuted in the same hardware environment, the primary claim made in SoftDist\nis no longer supported.\n", "date": "2024-06-11", "categories": []} {"id": "2406.09214", "title": "Applying Multi-Agent Negotiation to Solve the Production Routing Problem\n With Privacy Preserving", "abstract": " This paper presents a novel approach to address the Production Routing\nProblem with Privacy Preserving (PRPPP) in supply chain optimization. The\nintegrated optimization of production, inventory, distribution, and routing\ndecisions in real-world industry applications poses several challenges,\nincluding increased complexity, discrepancies between planning and execution,\nand constraints on information sharing. To mitigate these challenges, this\npaper proposes the use of intelligent agent negotiation within a hybrid\nMulti-Agent System (MAS) integrated with optimization algorithms. The MAS\nfacilitates communication and coordination among entities, encapsulates private\ninformation, and enables negotiation. This, along with optimization algorithms,\nmakes it a compelling framework for establishing optimal solutions. The\napproach is supported by real-world applications and synergies between MAS and\noptimization methods, demonstrating its effectiveness in addressing complex\nsupply chain optimization problems.\n", "date": "2024-06-13", "categories": []} {"id": "2406.07567", "title": "A GRASP-based memetic algorithm with path relinking for the far from\n most string problem", "abstract": " The FAR FROM MOST STRING PROBLEM (FFMSP) is a string selection problem. The\nobjective is to find a string whose distance to other strings in a certain\ninput set is above a given threshold for as many of those strings as possible.\nThis problem has links with some tasks in computational biology and its\nresolution has been shown to be very hard. We propose a memetic algorithm (MA)\nto tackle the FFMSP. This MA exploits a heuristic objective function for the\nproblem and features initialization of the population via a Greedy Randomized\nAdaptive Search Procedure (GRASP) metaheuristic, intensive recombination via\npath relinking and local improvement via hill climbing. An extensive empirical\nevaluation using problem instances of both random and biological origin is done\nto assess parameter sensitivity and draw performance comparisons with other\nstate-of-the-art techniques. The MA is shown to perform better than these\nlatter techniques with statistical significance.\n", "date": "2024-05-27", "categories": []} {"id": "2406.07573", "title": "Investigating the Potential of Using Large Language Models for\n Scheduling", "abstract": " The inaugural ACM International Conference on AI-powered Software introduced\nthe AIware Challenge, prompting researchers to explore AI-driven tools for\noptimizing conference programs through constrained optimization. We investigate\nthe use of Large Language Models (LLMs) for program scheduling, focusing on\nzero-shot learning and integer programming to measure paper similarity. Our\nstudy reveals that LLMs, even under zero-shot settings, create reasonably good\nfirst drafts of conference schedules. When clustering papers, using only titles\nas LLM inputs produces results closer to human categorization than using titles\nand abstracts with TFIDF. The code has been made publicly available.\n", "date": "2024-06-04", "categories": []} {"id": "2406.07579", "title": "GFPack++: Improving 2D Irregular Packing by Learning Gradient Field with\n Attention", "abstract": " 2D irregular packing is a classic combinatorial optimization problem with\nvarious applications, such as material utilization and texture atlas\ngeneration. This NP-hard problem requires efficient algorithms to optimize\nspace utilization. Conventional numerical methods suffer from slow convergence\nand high computational cost. Existing learning-based methods, such as the\nscore-based diffusion model, also have limitations, such as no rotation\nsupport, frequent collisions, and poor adaptability to arbitrary boundaries,\nand slow inferring. The difficulty of learning from teacher packing is to\ncapture the complex geometric relationships among packing examples, which\ninclude the spatial (position, orientation) relationships of objects, their\ngeometric features, and container boundary conditions. Representing these\nrelationships in latent space is challenging. We propose GFPack++, an\nattention-based gradient field learning approach that addresses this challenge.\nIt consists of two pivotal strategies: \\emph{attention-based geometry encoding}\nfor effective feature encoding and \\emph{attention-based relation encoding} for\nlearning complex relationships. We investigate the utilization distribution\nbetween the teacher and inference data and design a weighting function to\nprioritize tighter teacher data during training, enhancing learning\neffectiveness. Our diffusion model supports continuous rotation and outperforms\nexisting methods on various datasets. We achieve higher space utilization over\nseveral widely used baselines, one-order faster than the previous\ndiffusion-based method, and promising generalization for arbitrary boundaries.\nWe plan to release our source code and datasets to support further research in\nthis direction.\n", "date": "2024-06-09", "categories": []} {"id": "2406.07666", "title": "A Unified Framework for Integer Programming Formulation of Graph\n Matching Problems", "abstract": " Graph theory has been a powerful tool in solving difficult and complex\nproblems arising in all disciplines. In particular, graph matching is a\nclassical problem in pattern analysis with enormous applications. Many graph\nproblems have been formulated as a mathematical program and then solved using\nexact, heuristic, and/or approximated-guaranteed procedures. On the other hand,\ngraph theory has been a powerful tool in visualizing and understanding complex\nmathematical programming problems, especially integer programs. Formulating a\ngraph problem as a natural integer program (IP) is often a challenging task.\nHowever, an IP formulation of the problem has many advantages. Several\nresearchers have noted the need for natural IP formulation of graph theoretic\nproblems. The present study aims to provide a unified framework for IP\nformulation of graph-matching problems. Although there are many surveys on\ngraph matching problems, none is concerned with IP formulation. This paper is\nthe first to provide a comprehensive IP formulation for such problems. The\nframework includes a variety of graph optimization problems in the literature.\nWhile these problems have been studied by different research communities,\nhowever, the framework presented here helps to bring efforts from different\ndisciplines to tackle such diverse and complex problems. We hope the present\nstudy can significantly help to simplify some of the difficult problems arising\nin practice, especially in pattern analysis.\n", "date": "2024-06-11", "categories": []} {"id": "2307.12136", "title": "Using Reinforcement Learning for the Three-Dimensional Loading\n Capacitated Vehicle Routing Problem", "abstract": " Heavy goods vehicles are vital backbones of the supply chain delivery system\nbut also contribute significantly to carbon emissions with only 60% loading\nefficiency in the United Kingdom. Collaborative vehicle routing has been\nproposed as a solution to increase efficiency, but challenges remain to make\nthis a possibility. One key challenge is the efficient computation of viable\nsolutions for co-loading and routing. Current operations research methods\nsuffer from non-linear scaling with increasing problem size and are therefore\nbound to limited geographic areas to compute results in time for day-to-day\noperations. This only allows for local optima in routing and leaves global\noptimisation potential untouched. We develop a reinforcement learning model to\nsolve the three-dimensional loading capacitated vehicle routing problem in\napproximately linear time. While this problem has been studied extensively in\noperations research, no publications on solving it with reinforcement learning\nexist. We demonstrate the favourable scaling of our reinforcement learning\nmodel and benchmark our routing performance against state-of-the-art methods.\nThe model performs within an average gap of 3.83% to 8.10% compared to\nestablished methods. Our model not only represents a promising first step\ntowards large-scale logistics optimisation with reinforcement learning but also\nlays the foundation for this research stream. GitHub:\nhttps://github.com/if-loops/3L-CVRP\n", "date": "2023-07-22", "categories": []} {"id": "2406.06593", "title": "Differentiable Combinatorial Scheduling at Scale", "abstract": " This paper addresses the complex issue of resource-constrained scheduling, an\nNP-hard problem that spans critical areas including chip design and\nhigh-performance computing. Traditional scheduling methods often stumble over\nscalability and applicability challenges. We propose a novel approach using a\ndifferentiable combinatorial scheduling framework, utilizing Gumbel-Softmax\ndifferentiable sampling technique. This new technical allows for a fully\ndifferentiable formulation of linear programming (LP) based scheduling,\nextending its application to a broader range of LP formulations. To encode\ninequality constraints for scheduling tasks, we introduce \\textit{constrained\nGumbel Trick}, which adeptly encodes arbitrary inequality constraints.\nConsequently, our method facilitates an efficient and scalable scheduling via\ngradient descent without the need for training data. Comparative evaluations on\nboth synthetic and real-world benchmarks highlight our capability to\nsignificantly improve the optimization efficiency of scheduling, surpassing\nstate-of-the-art solutions offered by commercial and open-source solvers such\nas CPLEX, Gurobi, and CP-SAT in the majority of the designs.\n", "date": "2024-06-06", "categories": []} {"id": "2406.07325", "title": "Beyond Training: Optimizing Reinforcement Learning Based Job Shop\n Scheduling Through Adaptive Action Sampling", "abstract": " Learned construction heuristics for scheduling problems have become\nincreasingly competitive with established solvers and heuristics in recent\nyears. In particular, significant improvements have been observed in solution\napproaches using deep reinforcement learning (DRL). While much attention has\nbeen paid to the design of network architectures and training algorithms to\nachieve state-of-the-art results, little research has investigated the optimal\nuse of trained DRL agents during inference. Our work is based on the hypothesis\nthat, similar to search algorithms, the utilization of trained DRL agents\nshould be dependent on the acceptable computational budget. We propose a simple\nyet effective parameterization, called $\\delta$-sampling that manipulates the\ntrained action vector to bias agent behavior towards exploration or\nexploitation during solution construction. By following this approach, we can\nachieve a more comprehensive coverage of the search space while still\ngenerating an acceptable number of solutions. In addition, we propose an\nalgorithm for obtaining the optimal parameterization for such a given number of\nsolutions and any given trained agent. Experiments extending existing training\nprotocols for job shop scheduling problems with our inference method validate\nour hypothesis and result in the expected improvements of the generated\nsolutions.\n", "date": "2024-06-11", "categories": []} {"id": "2406.02872", "title": "Combinatorial Optimization with Automated Graph Neural Networks", "abstract": " In recent years, graph neural networks (GNNs) have become increasingly\npopular for solving NP-hard combinatorial optimization (CO) problems, such as\nmaximum cut and maximum independent set. The core idea behind these methods is\nto represent a CO problem as a graph and then use GNNs to learn the node/graph\nembedding with combinatorial information. Although these methods have achieved\npromising results, given a specific CO problem, the design of GNN architectures\nstill requires heavy manual work with domain knowledge. Existing automated GNNs\nare mostly focused on traditional graph learning problems, which is\ninapplicable to solving NP-hard CO problems. To this end, we present a new\nclass of \\textbf{AUTO}mated \\textbf{G}NNs for solving \\textbf{NP}-hard\nproblems, namely \\textbf{AutoGNP}. We represent CO problems by GNNs and focus\non two specific problems, i.e., mixed integer linear programming and quadratic\nunconstrained binary optimization. The idea of AutoGNP is to use graph neural\narchitecture search algorithms to automatically find the best GNNs for a given\nNP-hard combinatorial optimization problem. Compared with existing graph neural\narchitecture search algorithms, AutoGNP utilizes two-hop operators in the\narchitecture search space. Moreover, AutoGNP utilizes simulated annealing and a\nstrict early stopping policy to avoid local optimal solutions. Empirical\nresults on benchmark combinatorial problems demonstrate the superiority of our\nproposed model.\n", "date": "2024-06-05", "categories": []} {"id": "2406.03647", "title": "Decision-focused Graph Neural Networks for Combinatorial Optimization", "abstract": " In recent years, there has been notable interest in investigating\ncombinatorial optimization (CO) problems by neural-based framework. An emerging\nstrategy to tackle these challenging problems involves the adoption of graph\nneural networks (GNNs) as an alternative to traditional algorithms, a subject\nthat has attracted considerable attention. Despite the growing popularity of\nGNNs and traditional algorithm solvers in the realm of CO, there is limited\nresearch on their integrated use and the correlation between them within an\nend-to-end framework. The primary focus of our work is to formulate a more\nefficient and precise framework for CO by employing decision-focused learning\non graphs. Additionally, we introduce a decision-focused framework that\nutilizes GNNs to address CO problems with auxiliary support. To realize an\nend-to-end approach, we have designed two cascaded modules: (a) an unsupervised\ntrained graph predictive model, and (b) a solver for quadratic binary\nunconstrained optimization. Empirical evaluations are conducted on various\nclassical tasks, including maximum cut, maximum independent set, and minimum\nvertex cover. The experimental results on classical CO problems (i.e. MaxCut,\nMIS, and MVC) demonstrate the superiority of our method over both the\nstandalone GNN approach and classical methods.\n", "date": "2024-06-05", "categories": []} {"id": "2406.05830", "title": "Probabilistic Approach to Black-Box Binary Optimization with Budget\n Constraints: Application to Sensor Placement", "abstract": " We present a fully probabilistic approach for solving binary optimization\nproblems with black-box objective functions and with budget constraints. In the\nprobabilistic approach, the optimization variable is viewed as a random\nvariable and is associated with a parametric probability distribution. The\noriginal optimization problem is replaced with an optimization over the\nexpected value of the original objective, which is then optimized over the\nprobability distribution parameters. The resulting optimal parameter (optimal\npolicy) is used to sample the binary space to produce estimates of the optimal\nsolution(s) of the original binary optimization problem. The probability\ndistribution is chosen from the family of Bernoulli models because the\noptimization variable is binary. The optimization constraints generally\nrestrict the feasibility region. This can be achieved by modeling the random\nvariable with a conditional distribution given satisfiability of the\nconstraints. Thus, in this work we develop conditional Bernoulli distributions\nto model the random variable conditioned by the total number of nonzero\nentries, that is, the budget constraint. This approach (a) is generally\napplicable to binary optimization problems with nonstochastic black-box\nobjective functions and budget constraints; (b) accounts for budget constraints\nby employing conditional probabilities that sample only the feasible region and\nthus considerably reduces the computational cost compared with employing soft\nconstraints; and (c) does not employ soft constraints and thus does not require\ntuning of a regularization parameter, for example to promote sparsity, which is\nchallenging in sensor placement optimization problems. The proposed approach is\nverified numerically by using an idealized bilinear binary optimization problem\nand is validated by using a sensor placement experiment in a parameter\nidentification setup.\n", "date": "2024-06-09", "categories": []} {"id": "2405.17272", "title": "DPN: Decoupling Partition and Navigation for Neural Solvers of Min-max\n Vehicle Routing Problems", "abstract": " The min-max vehicle routing problem (min-max VRP) traverses all given\ncustomers by assigning several routes and aims to minimize the length of the\nlongest route. Recently, reinforcement learning (RL)-based sequential planning\nmethods have exhibited advantages in solving efficiency and optimality.\nHowever, these methods fail to exploit the problem-specific properties in\nlearning representations, resulting in less effective features for decoding\noptimal routes. This paper considers the sequential planning process of min-max\nVRPs as two coupled optimization tasks: customer partition for different routes\nand customer navigation in each route (i.e., partition and navigation). To\neffectively process min-max VRP instances, we present a novel attention-based\nPartition-and-Navigation encoder (P&N Encoder) that learns distinct embeddings\nfor partition and navigation. Furthermore, we utilize an inherent symmetry in\ndecoding routes and develop an effective agent-permutation-symmetric (APS) loss\nfunction. Experimental results demonstrate that the proposed\nDecoupling-Partition-Navigation (DPN) method significantly surpasses existing\nlearning-based methods in both single-depot and multi-depot min-max VRPs. Our\ncode is available at\n", "date": "2024-05-27", "categories": []} {"id": "2406.03503", "title": "Position: Rethinking Post-Hoc Search-Based Neural Approaches for Solving\n Large-Scale Traveling Salesman Problems", "abstract": " Recent advancements in solving large-scale traveling salesman problems (TSP)\nutilize the heatmap-guided Monte Carlo tree search (MCTS) paradigm, where\nmachine learning (ML) models generate heatmaps, indicating the probability\ndistribution of each edge being part of the optimal solution, to guide MCTS in\nsolution finding. However, our theoretical and experimental analysis raises\ndoubts about the effectiveness of ML-based heatmap generation. In support of\nthis, we demonstrate that a simple baseline method can outperform complex ML\napproaches in heatmap generation. Furthermore, we question the practical value\nof the heatmap-guided MCTS paradigm. To substantiate this, our findings show\nits inferiority to the LKH-3 heuristic despite the paradigm's reliance on\nproblem-specific, hand-crafted strategies. For the future, we suggest research\ndirections focused on developing more theoretically sound heatmap generation\nmethods and exploring autonomous, generalizable ML approaches for combinatorial\nproblems. The code is available for review:\nhttps://github.com/xyfffff/rethink_mcts_for_tsp.\n", "date": "2024-06-02", "categories": []} {"id": "2310.00112", "title": "Reinforcement Learning for Node Selection in Branch-and-Bound", "abstract": " A big challenge in branch and bound lies in identifying the optimal node\nwithin the search tree from which to proceed. Current state-of-the-art\nselectors utilize either hand-crafted ensembles that automatically switch\nbetween naive sub-node selectors, or learned node selectors that rely on\nindividual node data. We propose a novel simulation technique that uses\nreinforcement learning (RL) while considering the entire tree state, rather\nthan just isolated nodes. To achieve this, we train a graph neural network that\nproduces a probability distribution based on the path from the model's root to\nits \"to-be-selected\" leaves. Modelling node-selection as a probability\ndistribution allows us to train the model using state-of-the-art RL techniques\nthat capture both intrinsic node-quality and node-evaluation costs. Our method\ninduces a high quality node selection policy on a set of varied and complex\nproblem sets, despite only being trained on specially designed, synthetic\ntravelling salesmen problem (TSP) instances. Using such a fixed pretrained\npolicy shows significant improvements on several benchmarks in optimality gap\nreductions and per-node efficiency under strict time constraints.\n", "date": "2023-09-29", "categories": []} {"id": "2402.17606", "title": "Learning Topological Representations with Bidirectional Graph Attention\n Network for Solving Job Shop Scheduling Problem", "abstract": " Existing learning-based methods for solving job shop scheduling problems\n(JSSP) usually use off-the-shelf GNN models tailored to undirected graphs and\nneglect the rich and meaningful topological structures of disjunctive graphs\n(DGs). This paper proposes the topology-aware bidirectional graph attention\nnetwork (TBGAT), a novel GNN architecture based on the attention mechanism, to\nembed the DG for solving JSSP in a local search framework. Specifically, TBGAT\nembeds the DG from a forward and a backward view, respectively, where the\nmessages are propagated by following the different topologies of the views and\naggregated via graph attention. Then, we propose a novel operator based on the\nmessage-passing mechanism to calculate the forward and backward topological\nsorts of the DG, which are the features for characterizing the topological\nstructures and exploited by our model. In addition, we theoretically and\nexperimentally show that TBGAT has linear computational complexity to the\nnumber of jobs and machines, respectively, strengthening our method's practical\nvalue. Besides, extensive experiments on five synthetic datasets and seven\nclassic benchmarks show that TBGAT achieves new SOTA results by outperforming a\nwide range of neural methods by a large margin. All the code and data are\npublicly available online at https://github.com/zcaicaros/TBGAT.\n", "date": "2024-02-27", "categories": []} {"id": "2406.02294", "title": "Smaller Batches, Bigger Gains? Investigating the Impact of Batch Sizes\n on Reinforcement Learning Based Real-World Production Scheduling", "abstract": " Production scheduling is an essential task in manufacturing, with\nReinforcement Learning (RL) emerging as a key solution. In a previous work, RL\nwas utilized to solve an extended permutation flow shop scheduling problem\n(PFSSP) for a real-world production line with two stages, linked by a central\nbuffer. The RL agent was trained to sequence equallysized product batches to\nminimize setup efforts and idle times. However, the substantial impact caused\nby varying the size of these product batches has not yet been explored. In this\nfollow-up study, we investigate the effects of varying batch sizes, exploring\nboth the quality of solutions and the training dynamics of the RL agent. The\nresults demonstrate that it is possible to methodically identify reasonable\nboundaries for the batch size. These boundaries are determined on one side by\nthe increasing sample complexity associated with smaller batch sizes, and on\nthe other side by the decreasing flexibility of the agent when dealing with\nlarger batch sizes. This provides the practitioner the ability to make an\ninformed decision regarding the selection of an appropriate batch size.\nMoreover, we introduce and investigate two new curriculum learning strategies\nto enable the training with small batch sizes. The findings of this work offer\nthe potential for application in several industrial use cases with comparable\nscheduling problems.\n", "date": "2024-06-04", "categories": []} {"id": "2402.00138", "title": "Decomposable Submodular Maximization in Federated Setting", "abstract": " Submodular functions, as well as the sub-class of decomposable submodular\nfunctions, and their optimization appear in a wide range of applications in\nmachine learning, recommendation systems, and welfare maximization. However,\noptimization of decomposable submodular functions with millions of component\nfunctions is computationally prohibitive. Furthermore, the component functions\nmay be private (they might represent user preference function, for example) and\ncannot be widely shared. To address these issues, we propose a {\\em federated\noptimization} setting for decomposable submodular optimization. In this\nsetting, clients have their own preference functions, and a weighted sum of\nthese preferences needs to be maximized. We implement the popular {\\em\ncontinuous greedy} algorithm in this setting where clients take parallel small\nlocal steps towards the local solution and then the local changes are\naggregated at a central server. To address the large number of clients, the\naggregation is performed only on a subsampled set. Further, the aggregation is\nperformed only intermittently between stretches of parallel local steps, which\nreduces communication cost significantly. We show that our federated algorithm\nis guaranteed to provide a good approximate solution, even in the presence of\nabove cost-cutting measures. Finally, we show how the federated setting can be\nincorporated in solving fundamental discrete submodular optimization problems\nsuch as Maximum Coverage and Facility Location.\n", "date": "2024-01-31", "categories": []} {"id": "2405.20951", "title": "Monte Carlo Tree Search Satellite Scheduling Under Cloud Cover\n Uncertainty", "abstract": " Efficient utilization of satellite resources in dynamic environments remains\na challenging problem in satellite scheduling. This paper addresses the\nmulti-satellite collection scheduling problem (m-SatCSP), aiming to optimize\ntask scheduling over a constellation of satellites under uncertain conditions\nsuch as cloud cover. Leveraging Monte Carlo Tree Search (MCTS), a stochastic\nsearch algorithm, two versions of MCTS are explored to schedule satellites\neffectively. Hyperparameter tuning is conducted to optimize the algorithm's\nperformance. Experimental results demonstrate the effectiveness of the MCTS\napproach, outperforming existing methods in both solution quality and\nefficiency. Comparative analysis against other scheduling algorithms showcases\ncompetitive performance, positioning MCTS as a promising solution for satellite\ntask scheduling in dynamic environments.\n", "date": "2024-05-31", "categories": []} {"id": "2405.19977", "title": "Consistent Submodular Maximization", "abstract": " Maximizing monotone submodular functions under cardinality constraints is a\nclassic optimization task with several applications in data mining and machine\nlearning. In this paper we study this problem in a dynamic environment with\nconsistency constraints: elements arrive in a streaming fashion and the goal is\nmaintaining a constant approximation to the optimal solution while having a\nstable solution (i.e., the number of changes between two consecutive solutions\nis bounded). We provide algorithms in this setting with different trade-offs\nbetween consistency and approximation quality. We also complement our\ntheoretical results with an experimental analysis showing the effectiveness of\nour algorithms in real-world instances.\n", "date": "2024-05-30", "categories": []} {"id": "2405.19184", "title": "Promoting Two-sided Fairness in Dynamic Vehicle Routing Problem", "abstract": " Dynamic Vehicle Routing Problem (DVRP), is an extension of the classic\nVehicle Routing Problem (VRP), which is a fundamental problem in logistics and\ntransportation. Typically, DVRPs involve two stakeholders: service providers\nthat deliver services to customers and customers who raise requests from\ndifferent locations. Many real-world applications can be formulated as DVRP\nsuch as ridesharing and non-compliance capture. Apart from original objectives\nlike optimising total utility or efficiency, DVRP should also consider fairness\nfor all parties. Unfairness can induce service providers and customers to give\nup on the systems, leading to negative financial and social impacts. However,\nmost existing DVRP-related applications focus on improving fairness from a\nsingle side, and there have been few works considering two-sided fairness and\nutility optimisation concurrently. To this end, we propose a novel framework, a\nTwo-sided Fairness-aware Genetic Algorithm (named 2FairGA), which expands the\ngenetic algorithm from the original objective solely focusing on utility to\nmulti-objectives that incorporate two-sided fairness. Subsequently, the impact\nof injecting two fairness definitions into the utility-focused model and the\ncorrelation between any pair of the three objectives are explored. Extensive\nexperiments demonstrate the superiority of our proposed framework compared to\nthe state-of-the-art.\n", "date": "2024-05-29", "categories": []} {"id": "2311.13569", "title": "Combinatorial Optimization with Policy Adaptation using Latent Space\n Search", "abstract": " Combinatorial Optimization underpins many real-world applications and yet,\ndesigning performant algorithms to solve these complex, typically NP-hard,\nproblems remains a significant research challenge. Reinforcement Learning (RL)\nprovides a versatile framework for designing heuristics across a broad spectrum\nof problem domains. However, despite notable progress, RL has not yet\nsupplanted industrial solvers as the go-to solution. Current approaches\nemphasize pre-training heuristics that construct solutions but often rely on\nsearch procedures with limited variance, such as stochastically sampling\nnumerous solutions from a single policy or employing computationally expensive\nfine-tuning of the policy on individual problem instances. Building on the\nintuition that performant search at inference time should be anticipated during\npre-training, we propose COMPASS, a novel RL approach that parameterizes a\ndistribution of diverse and specialized policies conditioned on a continuous\nlatent space. We evaluate COMPASS across three canonical problems - Travelling\nSalesman, Capacitated Vehicle Routing, and Job-Shop Scheduling - and\ndemonstrate that our search strategy (i) outperforms state-of-the-art\napproaches on 11 standard benchmarking tasks and (ii) generalizes better,\nsurpassing all other approaches on a set of 18 procedurally transformed\ninstance distributions.\n", "date": "2023-11-13", "categories": []} {"id": "2405.17576", "title": "Container pre-marshalling problem minimizing CV@R under uncertainty of\n ship arrival times", "abstract": " This paper is concerned with the container pre-marshalling problem, which\ninvolves relocating containers in the storage area so that they can be\nefficiently loaded onto ships without reshuffles. In reality, however, ship\narrival times are affected by various external factors, which can cause the\norder of container retrieval to be different from the initial plan. To\nrepresent such uncertainty, we generate multiple scenarios from a multivariate\nprobability distribution of ship arrival times. We derive a mixed-integer\nlinear optimization model to find an optimal container layout such that the\nconditional value-at-risk is minimized for the number of misplaced containers\nresponsible for reshuffles. Moreover, we devise an exact algorithm based on the\ncutting-plane method to handle large-scale problems. Numerical experiments\nusing synthetic datasets demonstrate that our method can produce high-quality\ncontainer layouts compared with the conventional robust optimization model.\nAdditionally, our algorithm can speed up the computation of solving large-scale\nproblems.\n", "date": "2024-05-27", "categories": []} {"id": "2405.17676", "title": "Utilising a Quantum Hybrid Solver for Bi-objective Quadratic Assignment\n Problems", "abstract": " The intersection between quantum computing and optimisation has been an area\nof interest in recent years. There have been numerous studies exploring the\napplication of quantum and quantum-hybrid solvers to various optimisation\nproblems. This work explores scalarisation methods within the context of\nsolving the bi-objective quadratic assignment problem using a quantum-hybrid\nsolver. We show results that are consistent with previous research on a\ndifferent Ising machine.\n", "date": "2024-05-27", "categories": []} {"id": "2207.00885", "title": "Reinforcement Learning Approaches for the Orienteering Problem with\n Stochastic and Dynamic Release Dates", "abstract": " In this paper, we study a sequential decision-making problem faced by\ne-commerce carriers related to when to send out a vehicle from the central\ndepot to serve customer requests, and in which order to provide the service,\nunder the assumption that the time at which parcels arrive at the depot is\nstochastic and dynamic. The objective is to maximize the expected number of\nparcels that can be delivered during service hours. We propose two\nreinforcement learning (RL) approaches for solving this problem. These\napproaches rely on a look-ahead strategy in which future release dates are\nsampled in a Monte-Carlo fashion and a batch approach is used to approximate\nfuture routes. Both RL approaches are based on value function approximation -\none combines it with a consensus function (VFA-CF) and the other one with a\ntwo-stage stochastic integer linear programming model (VFA-2S). VFA-CF and\nVFA-2S do not need extensive training as they are based on very few\nhyper-parameters and make good use of integer linear programming (ILP) and\nbranch-and-cut-based exact methods to improve the quality of decisions. We also\nestablish sufficient conditions for partial characterization of optimal policy\nand integrate them into VFA-CF/VFA-2S. In an empirical study, we conduct a\ncompetitive analysis using upper bounds with perfect information. We also show\nthat VFA-CF and VFA-2S greatly outperform alternative approaches that: 1) do\nnot rely on future information, or 2) are based on point estimation of future\ninformation, or 3) employ heuristics rather than exact methods, or 4) use exact\nevaluations of future rewards.\n", "date": "2022-07-02", "categories": []} {"id": "2312.11527", "title": "A Simulated Annealing-Based Multiobjective Optimization Algorithm for\n Minimum Weight Minimum Connected Dominating Set Problem", "abstract": " Minimum connected dominating set problem is an NP-hard combinatorial\noptimization problem in graph theory. Finding connected dominating set is of\nhigh interest in various domains such as wireless sensor networks, optical\nnetworks, and systems biology. Its weighted variant named minimum weight\nconnected dominating set is also useful in such applications. In this paper, we\npropose a simulated annealing algorithm based on a greedy heuristic for\ntackling a variant of the minimum connected dominating set problem and that by\nexploiting two objectives together namely the cardinality and the total weight\nof the connected dominating set. Experimental results compared to those\nobtained by a recent proposed research show the superiority of our approach.\n", "date": "2023-12-13", "categories": []} {"id": "2402.02317", "title": "INViT: A Generalizable Routing Problem Solver with Invariant Nested View\n Transformer", "abstract": " Recently, deep reinforcement learning has shown promising results for\nlearning fast heuristics to solve routing problems. Meanwhile, most of the\nsolvers suffer from generalizing to an unseen distribution or distributions\nwith different scales. To address this issue, we propose a novel architecture,\ncalled Invariant Nested View Transformer (INViT), which is designed to enforce\na nested design together with invariant views inside the encoders to promote\nthe generalizability of the learned solver. It applies a modified policy\ngradient algorithm enhanced with data augmentations. We demonstrate that the\nproposed INViT achieves a dominant generalization performance on both TSP and\nCVRP problems with various distributions and different problem scales.\n", "date": "2024-02-04", "categories": []} {"id": "2402.13380", "title": "Toward TransfORmers: Revolutionizing the Solution of Mixed Integer\n Programs with Transformers", "abstract": " In this study, we introduce an innovative deep learning framework that\nemploys a transformer model to address the challenges of mixed-integer\nprograms, specifically focusing on the Capacitated Lot Sizing Problem (CLSP).\nOur approach, to our knowledge, is the first to utilize transformers to predict\nthe binary variables of a mixed-integer programming (MIP) problem.\nSpecifically, our approach harnesses the encoder decoder transformer's ability\nto process sequential data, making it well-suited for predicting binary\nvariables indicating production setup decisions in each period of the CLSP.\nThis problem is inherently dynamic, and we need to handle sequential decision\nmaking under constraints. We present an efficient algorithm in which CLSP\nsolutions are learned through a transformer neural network. The proposed\npost-processed transformer algorithm surpasses the state-of-the-art solver,\nCPLEX and Long Short-Term Memory (LSTM) in solution time, optimal gap, and\npercent infeasibility over 240K benchmark CLSP instances tested. After the ML\nmodel is trained, conducting inference on the model, reduces the MIP into a\nlinear program (LP). This transforms the ML-based algorithm, combined with an\nLP solver, into a polynomial-time approximation algorithm to solve a well-known\nNP-Hard problem, with almost perfect solution quality.\n", "date": "2024-02-20", "categories": []} {"id": "2405.08424", "title": "Tackling Prevalent Conditions in Unsupervised Combinatorial\n Optimization: Cardinality, Minimum, Covering, and More", "abstract": " Combinatorial optimization (CO) is naturally discrete, making machine\nlearning based on differentiable optimization inapplicable. Karalias & Loukas\n(2020) adapted the probabilistic method to incorporate CO into differentiable\noptimization. Their work ignited the research on unsupervised learning for CO,\ncomposed of two main components: probabilistic objectives and derandomization.\nHowever, each component confronts unique challenges. First, deriving objectives\nunder various conditions (e.g., cardinality constraints and minimum) is\nnontrivial. Second, the derandomization process is underexplored, and the\nexisting derandomization methods are either random sampling or naive rounding.\nIn this work, we aim to tackle prevalent (i.e., commonly involved) conditions\nin unsupervised CO. First, we concretize the targets for objective construction\nand derandomization with theoretical justification. Then, for various\nconditions commonly involved in different CO problems, we derive nontrivial\nobjectives and derandomization to meet the targets. Finally, we apply the\nderivations to various CO problems. Via extensive experiments on synthetic and\nreal-world graphs, we validate the correctness of our derivations and show our\nempirical superiority w.r.t. both optimization quality and speed.\n", "date": "2024-05-14", "categories": []} {"id": "2405.14995", "title": "Lower Bound on the Greedy Approximation Ratio for Adaptive Submodular\n Cover", "abstract": " We show that the greedy algorithm for adaptive-submodular cover has\napproximation ratio at least 1.3*(1+ln Q). Moreover, the instance demonstrating\nthis gap has Q=1. So, it invalidates a prior result in the paper ``Adaptive\nSubmodularity: A New Approach to Active Learning and Stochastic Optimization''\nby Golovin-Krause, that claimed a (1+ln Q)^2 approximation ratio for the same\nalgorithm.\n", "date": "2024-05-23", "categories": []} {"id": "2405.15569", "title": "Randomized heuristic repair for large-scale multidimensional knapsack\n problem", "abstract": " The multidimensional knapsack problem (MKP) is an NP-hard combinatorial\noptimization problem whose solution is determining a subset of maximum total\nprofit items that do not violate capacity constraints. Due to its hardness,\nlarge-scale MKP instances are usually a target for metaheuristics, a context in\nwhich effective feasibility maintenance strategies are crucial. In 1998, Chu\nand Beasley proposed an effective heuristic repair that is still relevant for\nrecent metaheuristics. However, due to its deterministic nature, the diversity\nof solutions such heuristic provides is insufficient for long runs. As a\nresult, the search for new solutions ceases after a while. This paper proposes\nan efficiency-based randomization strategy for the heuristic repair that\nincreases the variability of the repaired solutions without deteriorating\nquality and improves the overall results.\n", "date": "2024-05-24", "categories": []} {"id": "2208.08351", "title": "Minimum Cost Adaptive Submodular Cover", "abstract": " Adaptive submodularity is a fundamental concept in stochastic optimization,\nwith numerous applications such as sensor placement, hypothesis identification\nand viral marketing. We consider the problem of minimum cost cover of\nadaptive-submodular functions, and provide a $4(1+\\ln Q)$-approximation\nalgorithm, where $Q$ is the goal value. In fact, we consider a significantly\nmore general objective of minimizing the $p^{th}$ moment of the coverage cost,\nand show that our algorithm simultaneously achieves a $(p+1)^{p+1}\\cdot (\\ln\nQ+1)^p$ approximation guarantee for all $p\\ge 1$. All our approximation ratios\nare best possible up to constant factors (assuming $P\\ne NP$). Moreover, our\nresults also extend to the setting where one wants to cover {\\em multiple}\nadaptive-submodular functions. Finally, we evaluate the empirical performance\nof our algorithm on instances of hypothesis identification.\n", "date": "2022-08-17", "categories": []} {"id": "2405.13947", "title": "Leader Reward for POMO-Based Neural Combinatorial Optimization", "abstract": " Deep neural networks based on reinforcement learning (RL) for solving\ncombinatorial optimization (CO) problems are developing rapidly and have shown\na tendency to approach or even outperform traditional solvers. However,\nexisting methods overlook an important distinction: CO problems differ from\nother traditional problems in that they focus solely on the optimal solution\nprovided by the model within a specific length of time, rather than considering\nthe overall quality of all solutions generated by the model. In this paper, we\npropose Leader Reward and apply it during two different training phases of the\nPolicy Optimization with Multiple Optima (POMO) model to enhance the model's\nability to generate optimal solutions. This approach is applicable to a variety\nof CO problems, such as the Traveling Salesman Problem (TSP), the Capacitated\nVehicle Routing Problem (CVRP), and the Flexible Flow Shop Problem (FFSP), but\nalso works well with other POMO-based models or inference phase's strategies.\nWe demonstrate that Leader Reward greatly improves the quality of the optimal\nsolutions generated by the model. Specifically, we reduce the POMO's gap to the\noptimum by more than 100 times on TSP100 with almost no additional\ncomputational overhead.\n", "date": "2024-05-22", "categories": []} {"id": "2405.13992", "title": "Learning Cut Generating Functions for Integer Programming", "abstract": " The branch-and-cut algorithm is the method of choice to solve large scale\ninteger programming problems in practice. A key ingredient of branch-and-cut is\nthe use of cutting planes which are derived constraints that reduce the search\nspace for an optimal solution. Selecting effective cutting planes to produce\nsmall branch-and-cut trees is a critical challenge in the branch-and-cut\nalgorithm. Recent advances have employed a data-driven approach to select\noptimal cutting planes from a parameterized family, aimed at reducing the\nbranch-and-bound tree size (in expectation) for a given distribution of integer\nprogramming instances. We extend this idea to the selection of the best cut\ngenerating function (CGF), which is a tool in the integer programming\nliterature for generating a wide variety of cutting planes that generalize the\nwell-known Gomory Mixed-Integer (GMI) cutting planes. We provide rigorous\nsample complexity bounds for the selection of an effective CGF from certain\nparameterized families that provably performs well for any specified\ndistribution on the problem instances. Our empirical results show that the\nselected CGF can outperform the GMI cuts for certain distributions.\nAdditionally, we explore the sample complexity of using neural networks for\ninstance-dependent CGF selection.\n", "date": "2024-05-22", "categories": []} {"id": "2405.13994", "title": "Practical $0.385$-Approximation for Submodular Maximization Subject to a\n Cardinality Constraint", "abstract": " Non-monotone constrained submodular maximization plays a crucial role in\nvarious machine learning applications. However, existing algorithms often\nstruggle with a trade-off between approximation guarantees and practical\nefficiency. The current state-of-the-art is a recent $0.401$-approximation\nalgorithm, but its computational complexity makes it highly impractical. The\nbest practical algorithms for the problem only guarantee $1/e$-approximation.\nIn this work, we present a novel algorithm for submodular maximization subject\nto a cardinality constraint that combines a guarantee of $0.385$-approximation\nwith a low and practical query complexity of $O(n+k^2)$. Furthermore, we\nevaluate the empirical performance of our algorithm in experiments based on\nvarious machine learning applications, including Movie Recommendation, Image\nSummarization, and more. These experiments demonstrate the efficacy of our\napproach.\n", "date": "2024-05-22", "categories": []} {"id": "2405.12237", "title": "EKM: An exact, polynomial-time algorithm for the $K$-medoids problem", "abstract": " The $K$-medoids problem is a challenging combinatorial clustering task,\nwidely used in data analysis applications. While numerous algorithms have been\nproposed to solve this problem, none of these are able to obtain an exact\n(globally optimal) solution for the problem in polynomial time. In this paper,\nwe present EKM: a novel algorithm for solving this problem exactly with\nworst-case $O\\left(N^{K+1}\\right)$ time complexity. EKM is developed according\nto recent advances in transformational programming and combinatorial\ngeneration, using formal program derivation steps. The derived algorithm is\nprovably correct by construction. We demonstrate the effectiveness of our\nalgorithm by comparing it against various approximate methods on numerous\nreal-world datasets. We show that the wall-clock run time of our algorithm\nmatches the worst-case time complexity analysis on synthetic datasets, clearly\noutperforming the exponential time complexity of benchmark branch-and-bound\nbased MIP solvers. To our knowledge, this is the first, rigorously-proven\npolynomial time, practical algorithm for this ubiquitous problem.\n", "date": "2024-05-16", "categories": []} {"id": "2405.12252", "title": "Enhanced Deterministic Approximation Algorithm for Non-monotone\n Submodular Maximization under Knapsack Constraint with Linear Query\n Complexity", "abstract": " In this work, we consider the Submodular Maximization under Knapsack (SMK)\nconstraint problem over the ground set of size $n$. The problem recently\nattracted a lot of attention due to its applications in various domains of\ncombination optimization, artificial intelligence, and machine learning. We\nimprove the approximation factor of the fastest deterministic algorithm from\n$6+\\epsilon$ to $5+\\epsilon$ while keeping the best query complexity of $O(n)$,\nwhere $\\epsilon >0$ is a constant parameter. Our technique is based on\noptimizing the performance of two components: the threshold greedy subroutine\nand the building of two disjoint sets as candidate solutions. Besides, by\ncarefully analyzing the cost of candidate solutions, we obtain a tighter\napproximation factor.\n", "date": "2024-05-20", "categories": []} {"id": "2405.12262", "title": "Prompt Learning for Generalized Vehicle Routing", "abstract": " Neural combinatorial optimization (NCO) is a promising learning-based\napproach to solving various vehicle routing problems without much manual\nalgorithm design. However, the current NCO methods mainly focus on the\nin-distribution performance, while the real-world problem instances usually\ncome from different distributions. A costly fine-tuning approach or generalized\nmodel retraining from scratch could be needed to tackle the out-of-distribution\ninstances. Unlike the existing methods, this work investigates an efficient\nprompt learning approach in NCO for cross-distribution adaptation. To be\nconcrete, we propose a novel prompt learning method to facilitate fast\nzero-shot adaptation of a pre-trained model to solve routing problem instances\nfrom different distributions. The proposed model learns a set of prompts among\nvarious distributions and then selects the best-matched one to prompt a\npre-trained attention model for each problem instance. Extensive experiments\nshow that the proposed prompt learning approach facilitates the fast adaptation\nof pre-trained routing models. It also outperforms existing generalized models\non both in-distribution prediction and zero-shot generalization to a diverse\nset of new tasks. Our code implementation is available online\nhttps://github.com/FeiLiu36/PromptVRP.\n", "date": "2024-05-20", "categories": []} {"id": "2405.12475", "title": "GASE: Graph Attention Sampling with Edges Fusion for Solving Vehicle\n Routing Problems", "abstract": " Learning-based methods have become increasingly popular for solving vehicle\nrouting problems due to their near-optimal performance and fast inference\nspeed. Among them, the combination of deep reinforcement learning and graph\nrepresentation allows for the abstraction of node topology structures and\nfeatures in an encoder-decoder style. Such an approach makes it possible to\nsolve routing problems end-to-end without needing complicated heuristic\noperators designed by domain experts. Existing research studies have been\nfocusing on novel encoding and decoding structures via various neural network\nmodels to enhance the node embedding representation. Despite the sophisticated\napproaches applied, there is a noticeable lack of consideration for the\ngraph-theoretic properties inherent to routing problems. Moreover, the\npotential ramifications of inter-nodal interactions on the decision-making\nefficacy of the models have not been adequately explored. To bridge this gap,\nwe propose an adaptive Graph Attention Sampling with the Edges Fusion framework\n(GASE),where nodes' embedding is determined through attention calculation from\ncertain highly correlated neighbourhoods and edges, utilizing a filtered\nadjacency matrix. In detail, the selections of particular neighbours and\nadjacency edges are led by a multi-head attention mechanism, contributing\ndirectly to the message passing and node embedding in graph attention sampling\nnetworks. Furthermore, we incorporate an adaptive actor-critic algorithm with\npolicy improvements to expedite the training convergence. We then conduct\ncomprehensive experiments against baseline methods on learning-based VRP tasks\nfrom different perspectives. Our proposed model outperforms the existing\nmethods by 2.08\\%-6.23\\% and shows stronger generalization ability, achieving\nstate-of-the-art performance on randomly generated instances and real-world\ndatasets.\n", "date": "2024-05-21", "categories": []} {"id": "2405.11305", "title": "Large Neighborhood Prioritized Search for Combinatorial Optimization\n with Answer Set Programming", "abstract": " We propose Large Neighborhood Prioritized Search (LNPS) for solving\ncombinatorial optimization problems in Answer Set Programming (ASP). LNPS is a\nmetaheuristic that starts with an initial solution and then iteratively tries\nto find better solutions by alternately destroying and prioritized searching\nfor a current solution. Due to the variability of neighborhoods, LNPS allows\nfor flexible search without strongly depending on the destroy operators. We\npresent an implementation of LNPS based on ASP. The resulting heulingo solver\ndemonstrates that LNPS can significantly enhance the solving performance of ASP\nfor optimization. Furthermore, we establish the competitiveness of our LNPS\napproach by empirically contrasting it to (adaptive) large neighborhood search.\n", "date": "2024-05-18", "categories": []} {"id": "2405.11729", "title": "Optimization of Worker Scheduling at Logistics Depots Using Genetic\n Algorithms and Simulated Annealing", "abstract": " This paper addresses the optimization of scheduling for workers at a\nlogistics depot using a combination of genetic algorithm and simulated\nannealing algorithm. The efficient scheduling of permanent and temporary\nworkers is crucial for optimizing the efficiency of the logistics depot while\nminimizing labor usage. The study begins by establishing a 0-1 integer linear\nprogramming model, with decision variables determining the scheduling of\npermanent and temporary workers for each time slot on a given day. The\nobjective function aims to minimize person-days, while constraints ensure\nfulfillment of hourly labor requirements, limit workers to one time slot per\nday, cap consecutive working days for permanent workers, and maintain\nnon-negativity and integer constraints. The model is then solved using genetic\nalgorithms and simulated annealing. Results indicate that, for this problem,\ngenetic algorithms outperform simulated annealing in terms of solution quality.\nThe optimal solution reveals a minimum of 29857 person-days.\n", "date": "2024-05-20", "categories": []} {"id": "2207.03091", "title": "Online SuBmodular + SuPermodular (BP) Maximization with Bandit Feedback", "abstract": " In the context of online interactive machine learning with combinatorial\nobjectives, we extend purely submodular prior work to more general\nnon-submodular objectives. This includes: (1) those that are additively\ndecomposable into a sum of two terms (a monotone submodular and monotone\nsupermodular term, known as a BP decomposition); and (2) those that are only\nweakly submodular. In both cases, this allows representing not only competitive\n(submodular) but also complementary (supermodular) relationships between\nobjects, enhancing this setting to a broader range of applications (e.g., movie\nrecommendations, medical treatments, etc.) where this is beneficial. In the\ntwo-term case, moreover, we study not only the more typical monolithic feedback\napproach but also a novel framework where feedback is available separately for\neach term. With real-world practicality and scalability in mind, we integrate\nNystrom sketching techniques to significantly reduce the computational cost,\nincluding for the purely submodular case. In the Gaussian process contextual\nbandits setting, we show sub-linear theoretical regret bounds in all cases. We\nalso empirically show good applicability to recommendation systems and data\nsubset selection.\n", "date": "2022-07-07", "categories": []} {"id": "2309.16941", "title": "G4SATBench: Benchmarking and Advancing SAT Solving with Graph Neural\n Networks", "abstract": " Graph neural networks (GNNs) have recently emerged as a promising approach\nfor solving the Boolean Satisfiability Problem (SAT), offering potential\nalternatives to traditional backtracking or local search SAT solvers. However,\ndespite the growing volume of literature in this field, there remains a notable\nabsence of a unified dataset and a fair benchmark to evaluate and compare\nexisting approaches. To address this crucial gap, we present G4SATBench, the\nfirst benchmark study that establishes a comprehensive evaluation framework for\nGNN-based SAT solvers. In G4SATBench, we meticulously curate a large and\ndiverse set of SAT datasets comprising 7 problems with 3 difficulty levels and\nbenchmark a broad range of GNN models across various prediction tasks, training\nobjectives, and inference algorithms. To explore the learning abilities and\ncomprehend the strengths and limitations of GNN-based SAT solvers, we also\ncompare their solving processes with the heuristics in search-based SAT\nsolvers. Our empirical results provide valuable insights into the performance\nof GNN-based SAT solvers and further suggest that existing GNN models can\neffectively learn a solving strategy akin to greedy local search but struggle\nto learn backtracking search in the latent space. Our codebase is available at\nhttps://github.com/zhaoyu-li/G4SATBench.\n", "date": "2023-09-29", "categories": []} {"id": "2401.09556", "title": "Deep learning enhanced mixed integer optimization: Learning to reduce\n model dimensionality", "abstract": " This work introduces a framework to address the computational complexity\ninherent in Mixed-Integer Programming (MIP) models by harnessing the potential\nof deep learning. By employing deep learning, we construct problem-specific\nheuristics that identify and exploit common structures across MIP instances. We\ntrain deep learning models to estimate complicating binary variables for target\nMIP problem instances. The resulting reduced MIP models are solved using\nstandard off-the-shelf solvers. We present an algorithm for generating\nsynthetic data enhancing the robustness and generalizability of our models\nacross diverse MIP instances. We compare the effectiveness of (a) feed-forward\nneural networks (ANN) and (b) convolutional neural networks (CNN). To enhance\nthe framework's performance, we employ Bayesian optimization for hyperparameter\ntuning, aiming to maximize the occurrence of global optimum solutions. We apply\nthis framework to a flow-based facility location allocation MIP formulation\nthat describes long-term investment planning and medium-term tactical\nscheduling in a personalized medicine supply chain.\n", "date": "2024-01-17", "categories": []} {"id": "2402.00046", "title": "Introducing PetriRL: An Innovative Framework for JSSP Resolution\n Integrating Petri nets and Event-based Reinforcement Learning", "abstract": " Resource utilization and production process optimization are crucial for\ncompanies in today's competitive industrial landscape. Addressing the\ncomplexities of job shop scheduling problems (JSSP) is essential to improving\nproductivity, reducing costs, and ensuring timely delivery. We propose PetriRL,\na novel framework integrating Petri nets and deep reinforcement learning (DRL)\nfor JSSP optimization. PetriRL capitalizes on the inherent strengths of Petri\nnets in modelling discrete event systems while leveraging the advantages of a\ngraph structure. The Petri net governs automated components of the process,\nensuring adherence to JSSP constraints. This allows for synergistic\ncollaboration with optimization algorithms such as DRL, particularly in\ncritical decision-making. Unlike traditional methods, PetriRL eliminates the\nneed to preprocess JSSP instances into disjunctive graphs and enhances the\nexplainability of process status through its graphical structure based on\nplaces and transitions. Additionally, the inherent graph structure of Petri\nnets enables the dynamic additions of job operations during the inference phase\nwithout requiring agent retraining, thus enhancing flexibility. Experimental\nresults demonstrate PetriRL's robust generalization across various instance\nsizes and its competitive performance on public test benchmarks and randomly\ngenerated instances. Results are compared to a wide range of optimization\nsolutions such as heuristics, metaheuristics, and learning-based algorithms.\nFinally, the added values of the framework's key elements, such as event-based\ncontrol and action masking, are studied in the ablation study.\n", "date": "2024-01-23", "categories": []} {"id": "2405.04767", "title": "Test-Time Augmentation for Traveling Salesperson Problem", "abstract": " We propose Test-Time Augmentation (TTA) as an effective technique for\naddressing combinatorial optimization problems, including the Traveling\nSalesperson Problem. In general, deep learning models possessing the property\nof invariance, where the output is uniquely determined regardless of the node\nindices, have been proposed to learn graph structures efficiently. In contrast,\nwe interpret the permutation of node indices, which exchanges the elements of\nthe distance matrix, as a TTA scheme. The results demonstrate that our method\nis capable of obtaining shorter solutions than the latest models. Furthermore,\nwe show that the probability of finding a solution closer to an exact solution\nincreases depending on the augmentation size.\n", "date": "2024-05-08", "categories": []} {"id": "2308.14104", "title": "Towards Generalizable Neural Solvers for Vehicle Routing Problems via\n Ensemble with Transferrable Local Policy", "abstract": " Machine learning has been adapted to help solve NP-hard combinatorial\noptimization problems. One prevalent way is learning to construct solutions by\ndeep neural networks, which has been receiving more and more attention due to\nthe high efficiency and less requirement for expert knowledge. However, many\nneural construction methods for Vehicle Routing Problems~(VRPs) focus on\nsynthetic problem instances with specified node distributions and limited\nscales, leading to poor performance on real-world problems which usually\ninvolve complex and unknown node distributions together with large scales. To\nmake neural VRP solvers more practical, we design an auxiliary policy that\nlearns from the local transferable topological features, named local policy,\nand integrate it with a typical construction policy (which learns from the\nglobal information of VRP instances) to form an ensemble policy. With joint\ntraining, the aggregated policies perform cooperatively and complementarily to\nboost generalization. The experimental results on two well-known benchmarks,\nTSPLIB and CVRPLIB, of travelling salesman problem and capacitated VRP show\nthat the ensemble policy significantly improves both cross-distribution and\ncross-scale generalization performance, and even performs well on real-world\nproblems with several thousand nodes.\n", "date": "2023-08-27", "categories": []} {"id": "2405.01029", "title": "MVMoE: Multi-Task Vehicle Routing Solver with Mixture-of-Experts", "abstract": " Learning to solve vehicle routing problems (VRPs) has garnered much\nattention. However, most neural solvers are only structured and trained\nindependently on a specific problem, making them less generic and practical. In\nthis paper, we aim to develop a unified neural solver that can cope with a\nrange of VRP variants simultaneously. Specifically, we propose a multi-task\nvehicle routing solver with mixture-of-experts (MVMoE), which greatly enhances\nthe model capacity without a proportional increase in computation. We further\ndevelop a hierarchical gating mechanism for the MVMoE, delivering a good\ntrade-off between empirical performance and computational complexity.\nExperimentally, our method significantly promotes zero-shot generalization\nperformance on 10 unseen VRP variants, and showcases decent results on the\nfew-shot setting and real-world benchmark instances. We further conduct\nextensive studies on the effect of MoE configurations in solving VRPs, and\nobserve the superiority of hierarchical gating when facing out-of-distribution\ndata. The source code is available at:\nhttps://github.com/RoyalSkye/Routing-MVMoE.\n", "date": "2024-05-02", "categories": []} {"id": "2405.01997", "title": "Exploring Combinatorial Problem Solving with Large Language Models: A\n Case Study on the Travelling Salesman Problem Using GPT-3.5 Turbo", "abstract": " Large Language Models (LLMs) are deep learning models designed to generate\ntext based on textual input. Although researchers have been developing these\nmodels for more complex tasks such as code generation and general reasoning,\nfew efforts have explored how LLMs can be applied to combinatorial problems. In\nthis research, we investigate the potential of LLMs to solve the Travelling\nSalesman Problem (TSP). Utilizing GPT-3.5 Turbo, we conducted experiments\nemploying various approaches, including zero-shot in-context learning, few-shot\nin-context learning, and chain-of-thoughts (CoT). Consequently, we fine-tuned\nGPT-3.5 Turbo to solve a specific problem size and tested it using a set of\nvarious instance sizes. The fine-tuned models demonstrated promising\nperformance on problems identical in size to the training instances and\ngeneralized well to larger problems. Furthermore, to improve the performance of\nthe fine-tuned model without incurring additional training costs, we adopted a\nself-ensemble approach to improve the quality of the solutions.\n", "date": "2024-05-03", "categories": []} {"id": "2402.13608", "title": "Convergence Acceleration of Markov Chain Monte Carlo-based Gradient\n Descent by Deep Unfolding", "abstract": " This study proposes a trainable sampling-based solver for combinatorial\noptimization problems (COPs) using a deep-learning technique called deep\nunfolding. The proposed solver is based on the Ohzeki method that combines\nMarkov-chain Monte-Carlo (MCMC) and gradient descent, and its step sizes are\ntrained by minimizing a loss function. In the training process, we propose a\nsampling-based gradient estimation that substitutes auto-differentiation with a\nvariance estimation, thereby circumventing the failure of back propagation due\nto the non-differentiability of MCMC. The numerical results for a few COPs\ndemonstrated that the proposed solver significantly accelerated the convergence\nspeed compared with the original Ohzeki method.\n", "date": "2024-02-21", "categories": []} {"id": "2403.19561", "title": "Self-Improved Learning for Scalable Neural Combinatorial Optimization", "abstract": " The end-to-end neural combinatorial optimization (NCO) method shows promising\nperformance in solving complex combinatorial optimization problems without the\nneed for expert design. However, existing methods struggle with large-scale\nproblems, hindering their practical applicability. To overcome this limitation,\nthis work proposes a novel Self-Improved Learning (SIL) method for better\nscalability of neural combinatorial optimization. Specifically, we develop an\nefficient self-improved mechanism that enables direct model training on\nlarge-scale problem instances without any labeled data. Powered by an\ninnovative local reconstruction approach, this method can iteratively generate\nbetter solutions by itself as pseudo-labels to guide efficient model training.\nIn addition, we design a linear complexity attention mechanism for the model to\nefficiently handle large-scale combinatorial problem instances with low\ncomputation overhead. Comprehensive experiments on the Travelling Salesman\nProblem (TSP) and the Capacitated Vehicle Routing Problem (CVRP) with up to\n100K nodes in both uniform and real-world distributions demonstrate the\nsuperior scalability of our method.\n", "date": "2024-03-28", "categories": []} {"id": "2404.19518", "title": "MGCBS: An Optimal and Efficient Algorithm for Solving Multi-Goal\n Multi-Agent Path Finding Problem", "abstract": " With the expansion of the scale of robotics applications, the multi-goal\nmulti-agent pathfinding (MG-MAPF) problem began to gain widespread attention.\nThis problem requires each agent to visit pre-assigned multiple goal points at\nleast once without conflict. Some previous methods have been proposed to solve\nthe MG-MAPF problem based on Decoupling the goal Vertex visiting order search\nand the Single-agent pathfinding (DVS). However, this paper demonstrates that\nthe methods based on DVS cannot always obtain the optimal solution. To obtain\nthe optimal result, we propose the Multi-Goal Conflict-Based Search (MGCBS),\nwhich is based on Decoupling the goal Safe interval visiting order search and\nthe Single-agent pathfinding (DSS). Additionally, we present the\nTime-Interval-Space Forest (TIS Forest) to enhance the efficiency of MGCBS by\nmaintaining the shortest paths from any start point at any start time step to\neach safe interval at the goal points. The experiment demonstrates that our\nmethod can consistently obtain optimal results and execute up to 7 times faster\nthan the state-of-the-art method in our evaluation.\n", "date": "2024-04-30", "categories": []} {"id": "2307.04050", "title": "Optimization-based Learning for Dynamic Load Planning in Trucking\n Service Networks", "abstract": " The load planning problem is a critical challenge in service network design\nfor parcel carriers: it decides how many trailers to assign for dispatch over\ntime between pairs of terminals. Another key challenge is to determine a flow\nplan, which specifies how parcel volumes are assigned to planned loads. This\npaper considers the Outbound Load Planning Problem (OLPP) that considers flow\nand load planning challenges jointly in order to adjust loads and flows as the\ndemand forecast changes over time before the day of operations in a terminal.\nThe paper aims at developing a decision-support tool to inform planners making\nthese decisions at terminals across the network. The paper formulates the OLPP\nas a mixed-integer programming model and shows that it admits a large number of\nsymmetries in a network where each commodity can be routed through primary and\nalternate terminals. As a result, an optimization solver may return\nfundamentally different solutions to closely related problems, confusing\nplanners and reducing trust in optimization. To remedy this limitation, this\npaper proposes a lexicographical optimization approach that eliminates those\nsymmetries by generating optimal solutions staying close to a reference plan.\nMoreover, this paper designs an optimization proxy that addresses the\ncomputational challenges of the optimization model. The optimization proxy\ncombines a machine-learning model and a repair procedure to find near-optimal\nsolutions that satisfy real-time constraints imposed by planners in the loop.\nAn extensive computational study on industrial instances shows that the\noptimization proxy is orders of magnitude faster for generating solutions that\nare consistent with each other. The proposed approach also demonstrates the\nbenefits of the OLPP for load consolidation and the significant savings\nobtained from combining machine learning and optimization.\n", "date": "2023-07-08", "categories": []} {"id": "2404.17158", "title": "Online $\\mathrm{L}^{\\natural}$-Convex Minimization", "abstract": " An online decision-making problem is a learning problem in which a player\nrepeatedly makes decisions in order to minimize the long-term loss. These\nproblems that emerge in applications often have nonlinear combinatorial\nobjective functions, and developing algorithms for such problems has attracted\nconsiderable attention. An existing general framework for dealing with such\nobjective functions is the online submodular minimization. However, practical\nproblems are often out of the scope of this framework, since the domain of a\nsubmodular function is limited to a subset of the unit hypercube. To manage\nthis limitation of the existing framework, we in this paper introduce the\nonline $\\mathrm{L}^{\\natural}$-convex minimization, where an\n$\\mathrm{L}^{\\natural}$-convex function generalizes a submodular function so\nthat the domain is a subset of the integer lattice. We propose computationally\nefficient algorithms for the online $\\mathrm{L}^{\\natural}$-convex function\nminimization in two major settings: the full information and the bandit\nsettings. We analyze the regrets of these algorithms and show in particular\nthat our algorithm for the full information setting obtains a tight regret\nbound up to a constant factor. We also demonstrate several motivating examples\nthat illustrate the usefulness of the online $\\mathrm{L}^{\\natural}$-convex\nminimization.\n", "date": "2024-04-26", "categories": []} {"id": "2404.15852", "title": "QOPTLib: a Quantum Computing Oriented Benchmark for Combinatorial\n Optimization Problems", "abstract": " In this paper, we propose a quantum computing oriented benchmark for\ncombinatorial optimization. This benchmark, coined as QOPTLib, is composed of\n40 instances equally distributed over four well-known problems: Traveling\nSalesman Problem, Vehicle Routing Problem, one-dimensional Bin Packing Problem\nand the Maximum Cut Problem. The sizes of the instances in QOPTLib not only\ncorrespond to computationally addressable sizes, but also to the maximum length\napproachable with non-zero likelihood of getting a good result. In this regard,\nit is important to highlight that hybrid approaches are also taken into\nconsideration. Thus, this benchmark constitutes the first effort to provide\nusers a general-purpose dataset. Also in this paper, we introduce a first full\nsolving of QOPTLib using two solvers based on quantum annealing. Our main\nintention with this is to establish a preliminary baseline, hoping to inspire\nother researchers to beat these outcomes with newly proposed quantum-based\nalgorithms.\n", "date": "2024-04-24", "categories": []} {"id": "2404.14457", "title": "Graph Coloring Using Heat Diffusion", "abstract": " Graph coloring is a problem with varied applications in industry and science\nsuch as scheduling, resource allocation, and circuit design. The purpose of\nthis paper is to establish if a new gradient based iterative solver framework\nknown as heat diffusion can solve the graph coloring problem. We propose a\nsolution to the graph coloring problem using the heat diffusion framework. We\ncompare the solutions against popular methods and establish the competitiveness\nof heat diffusion method for the graph coloring problem.\n", "date": "2024-04-21", "categories": []} {"id": "2402.08772", "title": "Optimal Task Assignment and Path Planning using Conflict-Based Search\n with Precedence and Temporal Constraints", "abstract": " The Multi-Agent Path Finding (MAPF) problem entails finding collision-free\npaths for a set of agents, guiding them from their start to goal locations.\nHowever, MAPF does not account for several practical task-related constraints.\nFor example, agents may need to perform actions at goal locations with specific\nexecution times, adhering to predetermined orders and timeframes. Moreover,\ngoal assignments may not be predefined for agents, and the optimization\nobjective may lack an explicit definition. To incorporate task assignment, path\nplanning, and a user-defined objective into a coherent framework, this paper\nexamines the Task Assignment and Path Finding with Precedence and Temporal\nConstraints (TAPF-PTC) problem. We augment Conflict-Based Search (CBS) to\nsimultaneously generate task assignments and collision-free paths that adhere\nto precedence and temporal constraints, maximizing an objective quantified by\nthe return from a user-defined reward function in reinforcement learning (RL).\nExperimentally, we demonstrate that our algorithm, CBS-TA-PTC, can solve highly\nchallenging bomb-defusing tasks with precedence and temporal constraints\nefficiently relative to MARL and adapted Target Assignment and Path Finding\n(TAPF) methods.\n", "date": "2024-02-13", "categories": []} {"id": "2404.05223", "title": "ITA-ECBS: A Bounded-Suboptimal Algorithm for the Combined\n Target-Assignment and Path-Finding Problem", "abstract": " Multi-Agent Path Finding (MAPF), i.e., finding collision-free paths for\nmultiple robots, plays a critical role in many applications. Sometimes,\nassigning a target to each agent also presents a challenge. The Combined\nTarget-Assignment and Path-Finding (TAPF) problem, a variant of MAPF, requires\none to simultaneously assign targets to agents and plan collision-free paths\nfor agents. Several algorithms, including CBM, CBS-TA, and ITA-CBS, optimally\nsolve the TAPF problem, with ITA-CBS being the leading algorithm for minimizing\nflowtime. However, the only existing bounded-suboptimal algorithm ECBS-TA is\nderived from CBS-TA rather than ITA-CBS. So, it faces the same issues as\nCBS-TA, such as searching through multiple constraint trees and spending too\nmuch time on finding the next-best target assignment. We introduce ITA-ECBS,\nthe first bounded-suboptimal variant of ITA-CBS. Transforming ITA-CBS to its\nbounded-suboptimal variant is challenging because different constraint tree\nnodes can have different assignments of targets to agents. ITA-ECBS uses focal\nsearch to achieve efficiency and determines target assignments based on a new\nlower bound matrix. We show that it runs faster than ECBS-TA in 87.42% of\n54,033 test cases.\n", "date": "2024-04-08", "categories": []} {"id": "2404.13068", "title": "SmartPathfinder: Pushing the Limits of Heuristic Solutions for Vehicle\n Routing Problem with Drones Using Reinforcement Learning", "abstract": " The Vehicle Routing Problem with Drones (VRPD) seeks to optimize the routing\npaths for both trucks and drones, where the trucks are responsible for\ndelivering parcels to customer locations, and the drones are dispatched from\nthese trucks for parcel delivery, subsequently being retrieved by the trucks.\nGiven the NP-Hard complexity of VRPD, numerous heuristic approaches have been\nintroduced. However, improving solution quality and reducing computation time\nremain significant challenges. In this paper, we conduct a comprehensive\nexamination of heuristic methods designed for solving VRPD, distilling and\nstandardizing them into core elements. We then develop a novel reinforcement\nlearning (RL) framework that is seamlessly integrated with the heuristic\nsolution components, establishing a set of universal principles for\nincorporating the RL framework with heuristic strategies in an aim to improve\nboth the solution quality and computation speed. This integration has been\napplied to a state-of-the-art heuristic solution for VRPD, showcasing the\nsubstantial benefits of incorporating the RL framework. Our evaluation results\ndemonstrated that the heuristic solution incorporated with our RL framework not\nonly elevated the quality of solutions but also achieved rapid computation\nspeeds, especially when dealing with extensive customer locations.\n", "date": "2024-04-13", "categories": []} {"id": "2404.12638", "title": "Learning to Cut via Hierarchical Sequence/Set Model for Efficient\n Mixed-Integer Programming", "abstract": " Cutting planes (cuts) play an important role in solving mixed-integer linear\nprograms (MILPs), which formulate many important real-world applications. Cut\nselection heavily depends on (P1) which cuts to prefer and (P2) how many cuts\nto select. Although modern MILP solvers tackle (P1)-(P2) by human-designed\nheuristics, machine learning carries the potential to learn more effective\nheuristics. However, many existing learning-based methods learn which cuts to\nprefer, neglecting the importance of learning how many cuts to select.\nMoreover, we observe that (P3) what order of selected cuts to prefer\nsignificantly impacts the efficiency of MILP solvers as well. To address these\nchallenges, we propose a novel hierarchical sequence/set model (HEM) to learn\ncut selection policies. Specifically, HEM is a bi-level model: (1) a\nhigher-level module that learns how many cuts to select, (2) and a lower-level\nmodule -- that formulates the cut selection as a sequence/set to sequence\nlearning problem -- to learn policies selecting an ordered subset with the\ncardinality determined by the higher-level module. To the best of our\nknowledge, HEM is the first data-driven methodology that well tackles (P1)-(P3)\nsimultaneously. Experiments demonstrate that HEM significantly improves the\nefficiency of solving MILPs on eleven challenging MILP benchmarks, including\ntwo Huawei's real problems.\n", "date": "2024-04-19", "categories": []} {"id": "2303.14467", "title": "A Survey on the Densest Subgraph Problem and Its Variants", "abstract": " The Densest Subgraph Problem requires to find, in a given graph, a subset of\nvertices whose induced subgraph maximizes a measure of density. The problem has\nreceived a great deal of attention in the algorithmic literature since the\nearly 1970s, with many variants proposed and many applications built on top of\nthis basic definition. Recent years have witnessed a revival of research\ninterest in this problem with several important contributions, including some\ngroundbreaking results, published in 2022 and 2023. This survey provides a deep\noverview of the fundamental results and an exhaustive coverage of the many\nvariants proposed in the literature, with a special attention to the most\nrecent results. The survey also presents a comprehensive overview of\napplications and discusses some interesting open problems for this evergreen\nresearch topic.\n", "date": "2023-03-25", "categories": []} {"id": "2404.11907", "title": "Sampling-based Pareto Optimization for Chance-constrained Monotone\n Submodular Problems", "abstract": " Recently surrogate functions based on the tail inequalities were developed to\nevaluate the chance constraints in the context of evolutionary computation and\nseveral Pareto optimization algorithms using these surrogates were successfully\napplied in optimizing chance-constrained monotone submodular problems. However,\nthe difference in performance between algorithms using the surrogates and those\nemploying the direct sampling-based evaluation remains unclear. Within the\npaper, a sampling-based method is proposed to directly evaluate the chance\nconstraint. Furthermore, to address the problems with more challenging\nsettings, an enhanced GSEMO algorithm integrated with an adaptive sliding\nwindow, called ASW-GSEMO, is introduced. In the experiments, the ASW-GSEMO\nemploying the sampling-based approach is tested on the chance-constrained\nversion of the maximum coverage problem with different settings. Its results\nare compared with those from other algorithms using different surrogate\nfunctions. The experimental findings indicate that the ASW-GSEMO with the\nsampling-based evaluation approach outperforms other algorithms, highlighting\nthat the performances of algorithms using different evaluation methods are\ncomparable. Additionally, the behaviors of ASW-GSEMO are visualized to explain\nthe distinctions between it and the algorithms utilizing the surrogate\nfunctions.\n", "date": "2024-04-18", "categories": []} {"id": "2404.11458", "title": "Learn to Tour: Operator Design For Solution Feasibility Mapping in\n Pickup-and-delivery Traveling Salesman Problem", "abstract": " This paper aims to develop a learning method for a special class of traveling\nsalesman problems (TSP), namely, the pickup-and-delivery TSP (PDTSP), which\nfinds the shortest tour along a sequence of one-to-one pickup-and-delivery\nnodes. One-to-one here means that the transported people or goods are\nassociated with designated pairs of pickup and delivery nodes, in contrast to\nthat indistinguishable goods can be delivered to any nodes. In PDTSP,\nprecedence constraints need to be satisfied that each pickup node must be\nvisited before its corresponding delivery node. Classic operations research\n(OR) algorithms for PDTSP are difficult to scale to large-sized problems.\nRecently, reinforcement learning (RL) has been applied to TSPs. The basic idea\nis to explore and evaluate visiting sequences in a solution space. However,\nthis approach could be less computationally efficient, as it has to potentially\nevaluate many infeasible solutions of which precedence constraints are\nviolated. To restrict solution search within a feasible space, we utilize\noperators that always map one feasible solution to another, without spending\ntime exploring the infeasible solution space. Such operators are evaluated and\nselected as policies to solve PDTSPs in an RL framework. We make a comparison\nof our method and baselines, including classic OR algorithms and existing\nlearning methods. Results show that our approach can find tours shorter than\nbaselines.\n", "date": "2024-04-17", "categories": []} {"id": "2404.11606", "title": "Learning to Solve the Constrained Most Probable Explanation Task in\n Probabilistic Graphical Models", "abstract": " We propose a self-supervised learning approach for solving the following\nconstrained optimization task in log-linear models or Markov networks. Let $f$\nand $g$ be two log-linear models defined over the sets $\\mathbf{X}$ and\n$\\mathbf{Y}$ of random variables respectively. Given an assignment $\\mathbf{x}$\nto all variables in $\\mathbf{X}$ (evidence) and a real number $q$, the\nconstrained most-probable explanation (CMPE) task seeks to find an assignment\n$\\mathbf{y}$ to all variables in $\\mathbf{Y}$ such that $f(\\mathbf{x},\n\\mathbf{y})$ is maximized and $g(\\mathbf{x}, \\mathbf{y})\\leq q$. In our\nproposed self-supervised approach, given assignments $\\mathbf{x}$ to\n$\\mathbf{X}$ (data), we train a deep neural network that learns to output\nnear-optimal solutions to the CMPE problem without requiring access to any\npre-computed solutions. The key idea in our approach is to use first principles\nand approximate inference methods for CMPE to derive novel loss functions that\nseek to push infeasible solutions towards feasible ones and feasible solutions\ntowards optimal ones. We analyze the properties of our proposed method and\nexperimentally demonstrate its efficacy on several benchmark problems.\n", "date": "2024-04-17", "categories": []} {"id": "2402.16814", "title": "Box Facets and Cut Facets of Lifted Multicut Polytopes", "abstract": " The lifted multicut problem is a combinatorial optimization problem whose\nfeasible solutions relate one-to-one to the decompositions of a graph $G = (V,\nE)$. Given an augmentation $\\widehat{G} = (V, E \\cup F)$ of $G$ and given costs\n$c \\in \\mathbb{R}^{E \\cup F}$, the objective is to minimize the sum of those\n$c_{uw}$ with $uw \\in E \\cup F$ for which $u$ and $w$ are in distinct\ncomponents. For $F = \\emptyset$, the problem specializes to the multicut\nproblem, and for $E = \\tbinom{V}{2}$ to the clique partitioning problem. We\nstudy a binary linear program formulation of the lifted multicut problem. More\nspecifically, we contribute to the analysis of the associated lifted multicut\npolytopes: Firstly, we establish a necessary, sufficient and efficiently\ndecidable condition for a lower box inequality to define a facet. Secondly, we\nshow that deciding whether a cut inequality of the binary linear program\ndefines a facet is NP-hard.\n", "date": "2024-02-26", "categories": []} {"id": "2402.16891", "title": "Multi-Task Learning for Routing Problem with Cross-Problem Zero-Shot\n Generalization", "abstract": " Vehicle routing problems (VRPs), which can be found in numerous real-world\napplications, have been an important research topic for several decades.\nRecently, the neural combinatorial optimization (NCO) approach that leverages a\nlearning-based model to solve VRPs without manual algorithm design has gained\nsubstantial attention. However, current NCO methods typically require building\none model for each routing problem, which significantly hinders their practical\napplication for real-world industry problems with diverse attributes. In this\nwork, we make the first attempt to tackle the crucial challenge of\ncross-problem generalization. In particular, we formulate VRPs as different\ncombinations of a set of shared underlying attributes and solve them\nsimultaneously via a single model through attribute composition. In this way,\nour proposed model can successfully solve VRPs with unseen attribute\ncombinations in a zero-shot generalization manner. Extensive experiments are\nconducted on eleven VRP variants, benchmark datasets, and industry logistic\nscenarios. The results show that the unified model demonstrates superior\nperformance in the eleven VRPs, reducing the average gap to around 5% from over\n20% in the existing approach and achieving a significant performance boost on\nbenchmark datasets as well as a real-world logistics application. The source\ncode is included in https://github.com/FeiLiu36/MTNCO.\n", "date": "2024-02-23", "categories": []} {"id": "2303.10538", "title": "Unsupervised Learning for Solving the Travelling Salesman Problem", "abstract": " We propose UTSP, an unsupervised learning (UL) framework for solving the\nTravelling Salesman Problem (TSP). We train a Graph Neural Network (GNN) using\na surrogate loss. The GNN outputs a heat map representing the probability for\neach edge to be part of the optimal path. We then apply local search to\ngenerate our final prediction based on the heat map. Our loss function consists\nof two parts: one pushes the model to find the shortest path and the other\nserves as a surrogate for the constraint that the route should form a\nHamiltonian Cycle. Experimental results show that UTSP outperforms the existing\ndata-driven TSP heuristics. Our approach is parameter efficient as well as data\nefficient: the model takes $\\sim$ 10\\% of the number of parameters and $\\sim$\n0.2\\% of training samples compared with reinforcement learning or supervised\nlearning methods.\n", "date": "2023-03-19", "categories": []} {"id": "2404.06828", "title": "Proposed modified computational model for the amoeba-inspired\n combinatorial optimization machine", "abstract": " A single-celled amoeba can solve the traveling salesman problem through its\nshape-changing dynamics. In this paper, we examine roles of several elements in\na previously proposed computational model of the solution-search process of\namoeba and three modifications towards enhancing the solution-search\npreformance. We find that appropriate modifications can indeed significantly\nimprove the quality of solutions. It is also found that a condition associated\nwith the volume conservation can also be modified in contrast to the naive\nbelief that it is indispensable for the solution-search ability of amoeba. A\nproposed modified model shows much better performance.\n", "date": "2024-04-10", "categories": []} {"id": "2404.06014", "title": "Using 3-Objective Evolutionary Algorithms for the Dynamic Chance\n Constrained Knapsack Problem", "abstract": " Real-world optimization problems often involve stochastic and dynamic\ncomponents. Evolutionary algorithms are particularly effective in these\nscenarios, as they can easily adapt to uncertain and changing environments but\noften uncertainty and dynamic changes are studied in isolation. In this paper,\nwe explore the use of 3-objective evolutionary algorithms for the chance\nconstrained knapsack problem with dynamic constraints. In our setting, the\nweights of the items are stochastic and the knapsack's capacity changes over\ntime. We introduce a 3-objective formulation that is able to deal with the\nstochastic and dynamic components at the same time and is independent of the\nconfidence level required for the constraint. This new approach is then\ncompared to the 2-objective formulation which is limited to a single confidence\nlevel. We evaluate the approach using two different multi-objective\nevolutionary algorithms (MOEAs), namely the global simple evolutionary\nmulti-objective optimizer (GSEMO) and the multi-objective evolutionary\nalgorithm based on decomposition (MOEA/D), across various benchmark scenarios.\nOur analysis highlights the advantages of the 3-objective formulation over the\n2-objective formulation in addressing the dynamic chance constrained knapsack\nproblem.\n", "date": "2024-04-09", "categories": []} {"id": "2404.06447", "title": "The Central Spanning Tree Problem", "abstract": " Spanning trees are an important primitive in many data analysis tasks, when a\ndata set needs to be summarized in terms of its \"skeleton\", or when a\ntree-shaped graph over all observations is required for downstream processing.\nPopular definitions of spanning trees include the minimum spanning tree and the\noptimum distance spanning tree, a.k.a. the minimum routing cost tree. When\nsearching for the shortest spanning tree but admitting additional branching\npoints, even shorter spanning trees can be realized: Steiner trees.\nUnfortunately, both minimum spanning and Steiner trees are not robust with\nrespect to noise in the observations; that is, small perturbations of the\noriginal data set often lead to drastic changes in the associated spanning\ntrees. In response, we make two contributions when the data lies in a Euclidean\nspace: on the theoretical side, we introduce a new optimization problem, the\n\"(branched) central spanning tree\", which subsumes all previously mentioned\ndefinitions as special cases. On the practical side, we show empirically that\nthe (branched) central spanning tree is more robust to noise in the data, and\nas such is better suited to summarize a data set in terms of its skeleton. We\nalso propose a heuristic to address the NP-hard optimization problem, and\nillustrate its use on single cell RNA expression data from biology and 3D point\nclouds of plants.\n", "date": "2024-04-09", "categories": []} {"id": "2312.14213", "title": "A Reinforcement-Learning-Based Multiple-Column Selection Strategy for\n Column Generation", "abstract": " Column generation (CG) is one of the most successful approaches for solving\nlarge-scale linear programming (LP) problems. Given an LP with a prohibitively\nlarge number of variables (i.e., columns), the idea of CG is to explicitly\nconsider only a subset of columns and iteratively add potential columns to\nimprove the objective value. While adding the column with the most negative\nreduced cost can guarantee the convergence of CG, it has been shown that adding\nmultiple columns per iteration rather than a single column can lead to faster\nconvergence. However, it remains a challenge to design a multiple-column\nselection strategy to select the most promising columns from a large number of\ncandidate columns. In this paper, we propose a novel\nreinforcement-learning-based (RL) multiple-column selection strategy. To the\nbest of our knowledge, it is the first RL-based multiple-column selection\nstrategy for CG. The effectiveness of our approach is evaluated on two sets of\nproblems: the cutting stock problem and the graph coloring problem. Compared to\nseveral widely used single-column and multiple-column selection strategies, our\nRL-based multiple-column selection strategy leads to faster convergence and\nachieves remarkable reductions in the number of CG iterations and runtime.\n", "date": "2023-12-21", "categories": []} {"id": "2403.13421", "title": "Caching-Augmented Lifelong Multi-Agent Path Finding", "abstract": " Multi-Agent Path Finding (MAPF), which involves finding collision-free paths\nfor multiple robots, is crucial in various applications. Lifelong MAPF, where\ntargets are reassigned to agents as soon as they complete their initial\ntargets, offers a more accurate approximation of real-world warehouse planning.\nIn this paper, we present a novel mechanism named Caching-Augmented Lifelong\nMAPF (CAL-MAPF), designed to improve the performance of Lifelong MAPF. We have\ndeveloped a new type of map grid called cache for temporary item storage and\nreplacement, and created a locking mechanism to improve the planning solution's\nstability. A task assigner (TA) is designed for CAL-MAPF to allocate target\nlocations to agents and control agent status in different situations. CAL-MAPF\nhas been evaluated using various cache replacement policies and input task\ndistributions. We have identified three main factors significantly impacting\nCAL-MAPF performance through experimentation: suitable input task distribution,\nhigh cache hit rate, and smooth traffic. In general, CAL-MAPF has demonstrated\npotential for performance improvements in certain task distributions, map and\nagent configurations.\n", "date": "2024-03-20", "categories": []} {"id": "2404.02448", "title": "Electric Vehicle Routing Problem for Emergency Power Supply: Towards\n Telecom Base Station Relief", "abstract": " As a telecom provider, our company has a critical mission to maintain telecom\nservices even during power outages. To accomplish the mission, it is essential\nto maintain the power of the telecom base stations. Here we consider a solution\nwhere electric vehicles (EVs) directly supply power to base stations by\ntraveling to their locations. The goal is to find EV routes that minimize both\nthe total travel distance of all EVs and the number of downed base stations. In\nthis paper, we formulate this routing problem as a new variant of the Electric\nVehicle Routing Problem (EVRP) and propose a solver that combines a rule-based\nvehicle selector and a reinforcement learning (RL)-based node selector. The\nrule of the vehicle selector ensures the exact environmental states when the\nselected EV starts to move. In addition, the node selection by the RL model\nenables fast route generation, which is critical in emergencies. We evaluate\nour solver on both synthetic datasets and real datasets. The results show that\nour solver outperforms baselines in terms of the objective value and\ncomputation time. Moreover, we analyze the generalization and scalability of\nour solver, demonstrating the capability toward unseen settings and large-scale\nproblems. Check also our project page: https://ntt-dkiku.github.io/rl-evrpeps.\n", "date": "2024-04-03", "categories": []} {"id": "2404.04661", "title": "Transform then Explore: a Simple and Effective Technique for Exploratory\n Combinatorial Optimization with Reinforcement Learning", "abstract": " Many complex problems encountered in both production and daily life can be\nconceptualized as combinatorial optimization problems (COPs) over graphs.\nRecent years, reinforcement learning (RL) based models have emerged as a\npromising direction, which treat the COPs solving as a heuristic learning\nproblem. However, current finite-horizon-MDP based RL models have inherent\nlimitations. They are not allowed to explore adquately for improving solutions\nat test time, which may be necessary given the complexity of NP-hard\noptimization tasks. Some recent attempts solve this issue by focusing on reward\ndesign and state feature engineering, which are tedious and ad-hoc. In this\nwork, we instead propose a much simpler but more effective technique, named\ngauge transformation (GT). The technique is originated from physics, but is\nvery effective in enabling RL agents to explore to continuously improve the\nsolutions during test. Morever, GT is very simple, which can be implemented\nwith less than 10 lines of Python codes, and can be applied to a vast majority\nof RL models. Experimentally, we show that traditional RL models with GT\ntechnique produce the state-of-the-art performances on the MaxCut problem.\nFurthermore, since GT is independent of any RL models, it can be seamlessly\nintegrated into various RL frameworks, paving the way of these models for more\neffective explorations in the solving of general COPs.\n", "date": "2024-04-06", "categories": []} {"id": "2404.01308", "title": "Learning to Solve Job Shop Scheduling under Uncertainty", "abstract": " Job-Shop Scheduling Problem (JSSP) is a combinatorial optimization problem\nwhere tasks need to be scheduled on machines in order to minimize criteria such\nas makespan or delay. To address more realistic scenarios, we associate a\nprobability distribution with the duration of each task. Our objective is to\ngenerate a robust schedule, i.e. that minimizes the average makespan. This\npaper introduces a new approach that leverages Deep Reinforcement Learning\n(DRL) techniques to search for robust solutions, emphasizing JSSPs with\nuncertain durations. Key contributions of this research include: (1)\nadvancements in DRL applications to JSSPs, enhancing generalization and\nscalability, (2) a novel method for addressing JSSPs with uncertain durations.\nThe Wheatley approach, which integrates Graph Neural Networks (GNNs) and DRL,\nis made publicly available for further research and applications.\n", "date": "2024-03-04", "categories": []} {"id": "2404.00539", "title": "Solving the QAP by Two-Stage Graph Pointer Networks and Reinforcement\n Learning", "abstract": " Quadratic Assignment Problem (QAP) is a practical combinatorial optimization\nproblems that has been studied for several years. Since it is NP-hard, solving\nlarge problem instances of QAP is challenging. Although heuristics can find\nsemi-optimal solutions, the execution time significantly increases as the\nproblem size increases. Recently, solving combinatorial optimization problems\nby deep learning has been attracting attention as a faster solver than\nheuristics. Even with deep learning, however, solving large QAP is still\nchallenging. In this paper, we propose the deep reinforcement learning model\ncalled the two-stage graph pointer network (GPN) for solving QAP. Two-stage GPN\nrelies on GPN, which has been proposed for Euclidean Traveling Salesman Problem\n(TSP). First, we extend GPN for general TSP, and then we add new algorithms to\nthat model for solving QAP. Our experimental results show that our two-stage\nGPN provides semi-optimal solutions for benchmark problem instances from TSPlib\nand QAPLIB.\n", "date": "2024-03-31", "categories": []} {"id": "2308.08453", "title": "Tightest Admissible Shortest Path", "abstract": " The shortest path problem in graphs is fundamental to AI. Nearly all variants\nof the problem and relevant algorithms that solve them ignore edge-weight\ncomputation time and its common relation to weight uncertainty. This implies\nthat taking these factors into consideration can potentially lead to a\nperformance boost in relevant applications. Recently, a generalized framework\nfor weighted directed graphs was suggested, where edge-weight can be computed\n(estimated) multiple times, at increasing accuracy and run-time expense. We\nbuild on this framework to introduce the problem of finding the tightest\nadmissible shortest path (TASP); a path with the tightest suboptimality bound\non the optimal cost. This is a generalization of the shortest path problem to\nbounded uncertainty, where edge-weight uncertainty can be traded for\ncomputational cost. We present a complete algorithm for solving TASP, with\nguarantees on solution quality. Empirical evaluation supports the effectiveness\nof this approach.\n", "date": "2023-08-15", "categories": []} {"id": "2310.12387", "title": "Learning to Optimise Climate Sensor Placement using a Transformer", "abstract": " The optimal placement of sensors for environmental monitoring and disaster\nmanagement is a challenging problem due to its NP-hard nature. Traditional\nmethods for sensor placement involve exact, approximation, or heuristic\napproaches, with the latter being the most widely used. However, heuristic\nmethods are limited by expert intuition and experience. Deep learning (DL) has\nemerged as a promising approach for generating heuristic algorithms\nautomatically. In this paper, we introduce a novel sensor placement approach\nfocused on learning improvement heuristics using deep reinforcement learning\n(RL) methods. Our approach leverages an RL formulation for learning improvement\nheuristics, driven by an actor-critic algorithm for training the policy\nnetwork. We compare our method with several state-of-the-art approaches by\nconducting comprehensive experiments, demonstrating the effectiveness and\nsuperiority of our proposed approach in producing high-quality solutions. Our\nwork presents a promising direction for applying advanced DL and RL techniques\nto challenging climate sensor placement problems.\n", "date": "2023-10-18", "categories": []} {"id": "2403.17159", "title": "Less Is More -- On the Importance of Sparsification for Transformers and\n Graph Neural Networks for TSP", "abstract": " Most of the recent studies tackling routing problems like the Traveling\nSalesman Problem (TSP) with machine learning use a transformer or Graph Neural\nNetwork (GNN) based encoder architecture. However, many of them apply these\nencoders naively by allowing them to aggregate information over the whole TSP\ninstances. We, on the other hand, propose a data preprocessing method that\nallows the encoders to focus on the most relevant parts of the TSP instances\nonly. In particular, we propose graph sparsification for TSP graph\nrepresentations passed to GNNs and attention masking for TSP instances passed\nto transformers where the masks correspond to the adjacency matrices of the\nsparse TSP graph representations. Furthermore, we propose ensembles of\ndifferent sparsification levels allowing models to focus on the most promising\nparts while also allowing information flow between all nodes of a TSP instance.\nIn the experimental studies, we show that for GNNs appropriate sparsification\nand ensembles of different sparsification levels lead to substantial\nperformance increases of the overall architecture. We also design a new,\nstate-of-the-art transformer encoder with ensembles of attention masking. These\ntransformers increase model performance from a gap of $0.16\\%$ to $0.10\\%$ for\nTSP instances of size 100 and from $0.02\\%$ to $0.00\\%$ for TSP instances of\nsize 50.\n", "date": "2024-03-25", "categories": []} {"id": "2403.17108", "title": "Graph Protection under Multiple Simultaneous Attacks: A Heuristic\n Approach", "abstract": " This work focuses on developing an effective meta-heuristic approach to\nprotect against simultaneous attacks on nodes of a network modeled using a\ngraph. Specifically, we focus on the $k$-strong Roman domination problem, a\ngeneralization of the well-known Roman domination problem on graphs. This\ngeneral problem is about assigning integer weights to nodes that represent the\nnumber of field armies stationed at each node in order to satisfy the\nprotection constraints while minimizing the total weights. These constraints\nconcern the protection of a graph against any simultaneous attack consisting of\n$k \\in \\mathbb{N}$ nodes. An attack is considered repelled if each node labeled\n0 can be defended by borrowing an army from one of its neighboring nodes,\nensuring that the neighbor retains at least one army for self-defense. The\n$k$-SRD problem has practical applications in various areas, such as developing\ncounter-terrorism strategies or managing supply chain disruptions. The solution\nto this problem is notoriously difficult to find, as even checking the\nfeasibility of the proposed solution requires an exponential number of steps.\nWe propose a variable neighborhood search algorithm in which the feasibility of\nthe solution is checked by introducing the concept of quasi-feasibility, which\nis realized by careful sampling within the set of all possible attacks.\nExtensive experimental evaluations show the scalability and robustness of the\nproposed approach compared to the two exact approaches from the literature.\nExperiments are conducted with random networks from the literature and newly\nintroduced random wireless networks as well as with real-world networks. A\npractical application scenario, using real-world networks, involves applying\nour approach to graphs extracted from GeoJSON files containing geographic\nfeatures of hundreds of cities or larger regions.\n", "date": "2024-03-25", "categories": []} {"id": "2312.17272", "title": "Generating gradients in the energy landscape using rectified linear type\n cost functions for efficiently solving 0/1 matrix factorization in Simulated\n Annealing", "abstract": " The 0/1 matrix factorization defines matrix products using logical AND and OR\nas product-sum operators, revealing the factors influencing various decision\nprocesses. Instances and their characteristics are arranged in rows and\ncolumns. Formulating matrix factorization as an energy minimization problem and\nexploring it with Simulated Annealing (SA) theoretically enables finding a\nminimum solution in sufficient time. However, searching for the optimal\nsolution in practical time becomes problematic when the energy landscape has\nmany plateaus with flat slopes. In this work, we propose a method to facilitate\nthe solution process by applying a gradient to the energy landscape, using a\nrectified linear type cost function readily available in modern annealing\nmachines. We also propose a method to quickly obtain a solution by updating the\ncost function's gradient during the search process. Numerical experiments were\nconducted, confirming the method's effectiveness with both noise-free\nartificial and real data.\n", "date": "2023-12-27", "categories": []} {"id": "2403.14013", "title": "Towards a connection between the capacitated vehicle routing problem and\n the constrained centroid-based clustering", "abstract": " Efficiently solving a vehicle routing problem (VRP) in a practical runtime is\na critical challenge for delivery management companies. This paper explores\nboth a theoretical and experimental connection between the Capacitated Vehicle\nRouting Problem (CVRP) and the Constrained Centroid-Based Clustering (CCBC).\nReducing a CVRP to a CCBC is a synonym for a transition from an exponential to\na polynomial complexity using commonly known algorithms for clustering, i.e\nK-means. At the beginning, we conduct an exploratory analysis to highlight the\nexistence of such a relationship between the two problems through illustrative\nsmall-size examples and simultaneously deduce some mathematically-related\nformulations and properties. On a second level, the paper proposes a CCBC based\napproach endowed with some enhancements. The proposed framework consists of\nthree stages. At the first step, a constrained centroid-based clustering\nalgorithm generates feasible clusters of customers. This methodology\nincorporates three enhancement tools to achieve near-optimal clusters, namely:\na multi-start procedure for initial centroids, a customer assignment metric,\nand a self-adjustment mechanism for choosing the number of clusters. At the\nsecond step, a traveling salesman problem (T SP) solver is used to optimize the\norder of customers within each cluster. Finally, we introduce a process relying\non routes cutting and relinking procedure, which calls upon solving a linear\nand integer programming model to further improve the obtained routes. This step\nis inspired by the ruin & recreate algorithm. This approach is an extension of\nthe classical cluster-first, route-second method and provides near-optimal\nsolutions on well-known benchmark instances in terms of solution quality and\ncomputational runtime, offering a milestone in solving VRP.\n", "date": "2024-03-20", "categories": []} {"id": "2307.00618", "title": "Bounce: Reliable High-Dimensional Bayesian Optimization for\n Combinatorial and Mixed Spaces", "abstract": " Impactful applications such as materials discovery, hardware design, neural\narchitecture search, or portfolio optimization require optimizing\nhigh-dimensional black-box functions with mixed and combinatorial input spaces.\nWhile Bayesian optimization has recently made significant progress in solving\nsuch problems, an in-depth analysis reveals that the current state-of-the-art\nmethods are not reliable. Their performances degrade substantially when the\nunknown optima of the function do not have a certain structure. To fill the\nneed for a reliable algorithm for combinatorial and mixed spaces, this paper\nproposes Bounce that relies on a novel map of various variable types into\nnested embeddings of increasing dimensionality. Comprehensive experiments show\nthat Bounce reliably achieves and often even improves upon state-of-the-art\nperformance on a variety of high-dimensional problems.\n", "date": "2023-07-02", "categories": []} {"id": "2307.06742", "title": "Vehicle Dispatching and Routing of On-Demand Intercity Ride-Pooling\n Services: A Multi-Agent Hierarchical Reinforcement Learning Approach", "abstract": " The integrated development of city clusters has given rise to an increasing\ndemand for intercity travel. Intercity ride-pooling service exhibits\nconsiderable potential in upgrading traditional intercity bus services by\nimplementing demand-responsive enhancements. Nevertheless, its online\noperations suffer the inherent complexities due to the coupling of vehicle\nresource allocation among cities and pooled-ride vehicle routing. To tackle\nthese challenges, this study proposes a two-level framework designed to\nfacilitate online fleet management. Specifically, a novel multi-agent feudal\nreinforcement learning model is proposed at the upper level of the framework to\ncooperatively assign idle vehicles to different intercity lines, while the\nlower level updates the routes of vehicles using an adaptive large neighborhood\nsearch heuristic. Numerical studies based on the realistic dataset of Xiamen\nand its surrounding cities in China show that the proposed framework\neffectively mitigates the supply and demand imbalances, and achieves\nsignificant improvement in both the average daily system profit and order\nfulfillment ratio.\n", "date": "2023-07-13", "categories": []} {"id": "2403.08807", "title": "Effective anytime algorithm for multiobjective combinatorial\n optimization problems", "abstract": " In multiobjective optimization, the result of an optimization algorithm is a\nset of efficient solutions from which the decision maker selects one. It is\ncommon that not all the efficient solutions can be computed in a short time and\nthe search algorithm has to be stopped prematurely to analyze the solutions\nfound so far. A set of efficient solutions that are well-spread in the\nobjective space is preferred to provide the decision maker with a great variety\nof solutions. However, just a few exact algorithms in the literature exist with\nthe ability to provide such a well-spread set of solutions at any moment: we\ncall them anytime algorithms. We propose a new exact anytime algorithm for\nmultiobjective combinatorial optimization combining three novel ideas to\nenhance the anytime behavior. We compare the proposed algorithm with those in\nthe state-of-the-art for anytime multiobjective combinatorial optimization\nusing a set of 480 instances from different well-known benchmarks and four\ndifferent performance measures: the overall non-dominated vector generation\nratio, the hypervolume, the general spread and the additive epsilon indicator.\nA comprehensive experimental study reveals that our proposal outperforms the\nprevious algorithms in most of the instances.\n", "date": "2024-02-06", "categories": []} {"id": "2209.05208", "title": "Graph Neural Modeling of Network Flows", "abstract": " Network flow problems, which involve distributing traffic such that the\nunderlying infrastructure is used effectively, are ubiquitous in transportation\nand logistics. Among them, the general Multi-Commodity Network Flow (MCNF)\nproblem concerns the distribution of multiple flows of different sizes between\nseveral sources and sinks, while achieving effective utilization of the links.\nDue to the appeal of data-driven optimization, these problems have increasingly\nbeen approached using graph learning methods. In this paper, we propose a novel\ngraph learning architecture for network flow problems called Per-Edge Weights\n(PEW). This method builds on a Graph Attention Network and uses distinctly\nparametrized message functions along each link. We extensively evaluate the\nproposed solution through an Internet flow routing case study using $17$\nService Provider topologies and $2$ routing schemes. We show that PEW yields\nsubstantial gains over architectures whose global message function constrains\nthe routing unnecessarily. We also find that an MLP is competitive with other\nstandard architectures. Furthermore, we analyze the relationship between graph\nstructure and predictive performance for data-driven routing of flows, an\naspect that has not been considered by existing work in the area.\n", "date": "2022-09-12", "categories": []} {"id": "2312.07718", "title": "CaVE: A Cone-Aligned Approach for Fast Predict-then-optimize with Binary\n Linear Programs", "abstract": " The end-to-end predict-then-optimize framework, also known as\ndecision-focused learning, has gained popularity for its ability to integrate\noptimization into the training procedure of machine learning models that\npredict the unknown cost (objective function) coefficients of optimization\nproblems from contextual instance information. Naturally, most of the problems\nof interest in this space can be cast as integer linear programs. In this work,\nwe focus on binary linear programs (BLPs) and propose a new end-to-end training\nmethod to predict-then-optimize. Our method, Cone-aligned Vector Estimation\n(CaVE), aligns the predicted cost vectors with the normal cone corresponding to\nthe true optimal solution of a training instance. When the predicted cost\nvector lies inside the cone, the optimal solution to the linear relaxation of\nthe binary problem is optimal. This alignment not only produces decision-aware\nlearning models but also dramatically reduces training time as it circumvents\nthe need to solve BLPs to compute a loss function with its gradients.\nExperiments across multiple datasets show that our method exhibits a favorable\ntrade-off between training time and solution quality, particularly with\nlarge-scale optimization problems such as vehicle routing, a hard BLP that has\nyet to benefit from predict-then-optimize methods in the literature due to its\ndifficulty.\n", "date": "2023-12-12", "categories": []} {"id": "2401.16580", "title": "Attention-based Reinforcement Learning for Combinatorial Optimization:\n Application to Job Shop Scheduling Problem", "abstract": " Job shop scheduling problems represent a significant and complex facet of\ncombinatorial optimization problems, which have traditionally been addressed\nthrough either exact or approximate solution methodologies. However, the\npractical application of these solutions is often challenged due to the\ncomplexity of real-world problems. Even when utilizing an approximate solution\napproach, the time required to identify a near-optimal solution can be\nprohibitively extensive, and the solutions derived are generally not applicable\nto new problems. This study proposes an innovative attention-based\nreinforcement learning method specifically designed for the category of job\nshop scheduling problems. This method integrates a policy gradient\nreinforcement learning approach with a modified transformer architecture. A key\nfinding of this research is the ability of our trained learners within the\nproposed method to be repurposed for larger-scale problems that were not part\nof the initial training set. Furthermore, empirical evidence demonstrates that\nour approach surpasses the results of recent studies and outperforms commonly\nimplemented heuristic rules. This suggests that our method offers a promising\navenue for future research and practical application in the field of job shop\nscheduling problems.\n", "date": "2024-01-29", "categories": []} {"id": "2403.09742", "title": "A Short Review on Novel Approaches for Maximum Clique Problem: from\n Classical algorithms to Graph Neural Networks and Quantum algorithms", "abstract": " This manuscript provides a comprehensive review of the Maximum Clique\nProblem, a computational problem that involves finding subsets of vertices in a\ngraph that are all pairwise adjacent to each other. The manuscript covers in a\nsimple way classical algorithms for solving the problem and includes a review\nof recent developments in graph neural networks and quantum algorithms. The\nreview concludes with benchmarks for testing classical as well as new learning,\nand quantum algorithms.\n", "date": "2024-03-13", "categories": []} {"id": "2403.09925", "title": "Surrogate Assisted Monte Carlo Tree Search in Combinatorial Optimization", "abstract": " Industries frequently adjust their facilities network by opening new branches\nin promising areas and closing branches in areas where they expect low profits.\nIn this paper, we examine a particular class of facility location problems. Our\nobjective is to minimize the loss of sales resulting from the removal of\nseveral retail stores. However, estimating sales accurately is expensive and\ntime-consuming. To overcome this challenge, we leverage Monte Carlo Tree Search\n(MCTS) assisted by a surrogate model that computes evaluations faster. Results\nsuggest that MCTS supported by a fast surrogate function can generate solutions\nfaster while maintaining a consistent solution compared to MCTS that does not\nbenefit from the surrogate function.\n", "date": "2024-03-14", "categories": []} {"id": "2312.02544", "title": "Characterization of Locality in Spin States and Forced Moves for\n Optimizations", "abstract": " Ising formulations are widely utilized to solve combinatorial optimization\nproblems, and a variety of quantum or semiconductor-based hardware has recently\nbeen made available. In combinatorial optimization problems, the existence of\nlocal minima in energy landscapes is problematic to use to seek the global\nminimum. We note that the aim of the optimization is not to obtain exact\nsamplings from the Boltzmann distribution, and there is thus no need to satisfy\ndetailed balance conditions. In light of this fact, we develop an algorithm to\nget out of the local minima efficiently while it does not yield the exact\nsamplings. For this purpose, we utilize a feature that characterizes locality\nin the current state, which is easy to obtain with a type of specialized\nhardware. Furthermore, as the proposed algorithm is based on a rejection-free\nalgorithm, the computational cost is low. In this work, after presenting the\ndetails of the proposed algorithm, we report the results of numerical\nexperiments that demonstrate the effectiveness of the proposed feature and\nalgorithm.\n", "date": "2023-12-05", "categories": []} {"id": "2403.08839", "title": "Learning-Enhanced Neighborhood Selection for the Vehicle Routing Problem\n with Time Windows", "abstract": " Large Neighborhood Search (LNS) is a universal approach that is broadly\napplicable and has proven to be highly efficient in practice for solving\noptimization problems. We propose to integrate machine learning (ML) into LNS\nto assist in deciding which parts of the solution should be destroyed and\nrepaired in each iteration of LNS. We refer to our new approach as\nLearning-Enhanced Neighborhood Selection (LENS for short). Our approach is\nuniversally applicable, i.e., it can be applied to any LNS algorithm to amplify\nthe workings of the destroy algorithm. In this paper, we demonstrate the\npotential of LENS on the fundamental Vehicle Routing Problem with Time Windows\n(VRPTW). We implemented an LNS algorithm for VRPTW and collected data on\ngenerated novel training instances derived from well-known, extensively\nutilized benchmark datasets. We trained our LENS approach with this data and\ncompared the experimental results of our approach with two benchmark\nalgorithms: a random neighborhood selection method to show that LENS learns to\nmake informed choices and an oracle neighborhood selection method to\ndemonstrate the potential of our LENS approach. With LENS, we obtain results\nthat significantly improve the quality of the solutions.\n", "date": "2024-03-13", "categories": []} {"id": "2403.09249", "title": "Leveraging Constraint Programming in a Deep Learning Approach for\n Dynamically Solving the Flexible Job-Shop Scheduling Problem", "abstract": " Recent advancements in the flexible job-shop scheduling problem (FJSSP) are\nprimarily based on deep reinforcement learning (DRL) due to its ability to\ngenerate high-quality, real-time solutions. However, DRL approaches often fail\nto fully harness the strengths of existing techniques such as exact methods or\nconstraint programming (CP), which can excel at finding optimal or near-optimal\nsolutions for smaller instances. This paper aims to integrate CP within a deep\nlearning (DL) based methodology, leveraging the benefits of both. In this\npaper, we introduce a method that involves training a DL model using optimal\nsolutions generated by CP, ensuring the model learns from high-quality data,\nthereby eliminating the need for the extensive exploration typical in DRL and\nenhancing overall performance. Further, we integrate CP into our DL framework\nto jointly construct solutions, utilizing DL for the initial complex stages and\ntransitioning to CP for optimal resolution as the problem is simplified. Our\nhybrid approach has been extensively tested on three public FJSSP benchmarks,\ndemonstrating superior performance over five state-of-the-art DRL approaches\nand a widely-used CP solver. Additionally, with the objective of exploring the\napplication to other combinatorial optimization problems, promising preliminary\nresults are presented on applying our hybrid approach to the traveling salesman\nproblem, combining an exact method with a well-known DRL method.\n", "date": "2024-03-14", "categories": []} {"id": "2403.09361", "title": "A Multi-population Integrated Approach for Capacitated Location Routing", "abstract": " The capacitated location-routing problem involves determining the depots from\na set of candidate capacitated depot locations and finding the required routes\nfrom the selected depots to serve a set of customers whereas minimizing a cost\nfunction that includes the cost of opening the chosen depots, the fixed\nutilization cost per vehicle used, and the total cost (distance) of the routes.\nThis paper presents a multi-population integrated framework in which a\nmulti-depot edge assembly crossover generates promising offspring solutions\nfrom the perspective of both depot location and route edge assembly. The method\nincludes an effective neighborhood-based local search, a feasibility-restoring\nprocedure and a diversification-oriented mutation. Of particular interest is\nthe multi-population scheme which organizes the population into multiple\nsubpopulations based on depot configurations. Extensive experiments on 281\nbenchmark instances from the literature show that the algorithm performs\nremarkably well, by improving 101 best-known results (new upper bounds) and\nmatching 84 best-known results. Additional experiments are presented to gain\ninsight into the role of the key elements of the algorithm.\n", "date": "2024-03-14", "categories": []} {"id": "2403.06026", "title": "Towards a Generic Representation of Combinatorial Problems for\n Learning-Based Approaches", "abstract": " In recent years, there has been a growing interest in using learning-based\napproaches for solving combinatorial problems, either in an end-to-end manner\nor in conjunction with traditional optimization algorithms. In both scenarios,\nthe challenge lies in encoding the targeted combinatorial problems into a\nstructure compatible with the learning algorithm. Many existing works have\nproposed problem-specific representations, often in the form of a graph, to\nleverage the advantages of \\textit{graph neural networks}. However, these\napproaches lack generality, as the representation cannot be easily transferred\nfrom one combinatorial problem to another one. While some attempts have been\nmade to bridge this gap, they still offer a partial generality only. In\nresponse to this challenge, this paper advocates for progress toward a fully\ngeneric representation of combinatorial problems for learning-based approaches.\nThe approach we propose involves constructing a graph by breaking down any\nconstraint of a combinatorial problem into an abstract syntax tree and\nexpressing relationships (e.g., a variable involved in a constraint) through\nthe edges. Furthermore, we introduce a graph neural network architecture\ncapable of efficiently learning from this representation. The tool provided\noperates on combinatorial problems expressed in the XCSP3 format, handling all\nthe constraints available in the 2023 mini-track competition. Experimental\nresults on four combinatorial problems demonstrate that our architecture\nachieves performance comparable to dedicated architectures while maintaining\ngenerality. Our code and trained models are publicly available at\n\\url{https://github.com/corail-research/learning-generic-csp}.\n", "date": "2024-03-09", "categories": []} {"id": "2403.07886", "title": "A Memetic Algorithm To Find a Hamiltonian Cycle in a Hamiltonian Graph", "abstract": " We present a memetic algorithm (\\maa) approach for finding a Hamiltonian\ncycle in a Hamiltonian graph. The \\ma is based on a proven approach to the\nAsymmetric Travelling Salesman Problem (\\atspp) that, in this contribution, is\nboosted by the introduction of more powerful local searches. Our approach also\nintroduces a novel technique that sparsifies the input graph under\nconsideration for Hamiltonicity and dynamically augments it during the search.\nSuch a combined heuristic approach helps to prove Hamiltonicity by finding a\nHamiltonian cycle in less time. In addition, we also employ a recently\nintroduced polynomial-time reduction from the \\hamcyc to the Symmetric \\tsp,\nwhich is based on computing the transitive closure of the graph. Although our\napproach is a metaheuristic, i.e., it does not give a theoretical guarantee for\nfinding a Hamiltonian cycle, we have observed that the method is successful in\npractice in verifying the Hamiltonicity of a larger number of instances from\nthe \\textit{Flinder University Hamiltonian Cycle Problem Challenge Set}\n(\\fhcpsc), even for the graphs that have large treewidth. The experiments on\nthe \\fhcpscc instances and a computational comparison with five recent\nstate-of-the-art baseline approaches show that the proposed method outperforms\nthose for the majority of the instances in the \\fhcpsc.\n", "date": "2024-02-01", "categories": []} {"id": "2403.06995", "title": "Exact algorithms and heuristics for capacitated covering salesman\n problems", "abstract": " This paper introduces the Capacitated Covering Salesman Problem (CCSP),\napproaching the notion of service by coverage in capacitated vehicle routing\nproblems. In CCSP, locations where vehicles can transit are provided, some of\nwhich have customers with demands. The objective is to service customers\nthrough a fleet of vehicles based in a depot, minimizing the total distance\ntraversed by the vehicles. CCSP is unique in the sense that customers, to be\nserviced, do not need to be visited by a vehicle. Instead, they can be serviced\nif they are within a coverage area of the vehicle. This assumption is motivated\nby applications in which some customers are unreachable (e.g., forbidden access\nto vehicles) or visiting every customer is impractical. In this work,\noptimization methodologies are proposed for the CCSP based on ILP (Integer\nLinear Programming) and BRKGA (Biased Random-Key Genetic Algorithm)\nmetaheuristic. Computational experiments conducted on a benchmark of instances\nfor the CCSP evaluate the performance of the methodologies with respect to\nprimal bounds. Furthermore, our ILP formulation is extended in order to create\na novel MILP (Mixed Integer Linear Programming) for the Multi-Depot Covering\nTour Vehicle Routing Problem (MDCTVRP). Computational experiments show that the\nextended MILP formulation outperformed the previous state-of-the-art exact\napproach with respect to optimality gaps. In particular, optimal solutions were\nobtained for several previously unsolved instances.\n", "date": "2024-03-03", "categories": []} {"id": "2403.07028", "title": "An Efficient Learning-based Solver Comparable to Metaheuristics for the\n Capacitated Arc Routing Problem", "abstract": " Recently, neural networks (NN) have made great strides in combinatorial\noptimization. However, they face challenges when solving the capacitated arc\nrouting problem (CARP) which is to find the minimum-cost tour covering all\nrequired edges on a graph, while within capacity constraints. In tackling CARP,\nNN-based approaches tend to lag behind advanced metaheuristics, since they lack\ndirected arc modeling and efficient learning methods tailored for complex CARP.\nIn this paper, we introduce an NN-based solver to significantly narrow the gap\nwith advanced metaheuristics while exhibiting superior efficiency. First, we\npropose the direction-aware attention model (DaAM) to incorporate\ndirectionality into the embedding process, facilitating more effective\none-stage decision-making. Second, we design a supervised reinforcement\nlearning scheme that involves supervised pre-training to establish a robust\ninitial policy for subsequent reinforcement fine-tuning. It proves particularly\nvaluable for solving CARP that has a higher complexity than the node routing\nproblems (NRPs). Finally, a path optimization method is proposed to adjust the\ndepot return positions within the path generated by DaAM. Experiments\nillustrate that our approach surpasses heuristics and achieves decision quality\ncomparable to state-of-the-art metaheuristics for the first time while\nmaintaining superior efficiency.\n", "date": "2024-03-11", "categories": []} {"id": "2310.02807", "title": "A Deep Instance Generative Framework for MILP Solvers Under Limited Data\n Availability", "abstract": " In the past few years, there has been an explosive surge in the use of\nmachine learning (ML) techniques to address combinatorial optimization (CO)\nproblems, especially mixed-integer linear programs (MILPs). Despite the\nachievements, the limited availability of real-world instances often leads to\nsub-optimal decisions and biased solver assessments, which motivates a suite of\nsynthetic MILP instance generation techniques. However, existing methods either\nrely heavily on expert-designed formulations or struggle to capture the rich\nfeatures of real-world instances. To tackle this problem, we propose G2MILP,\nthe first deep generative framework for MILP instances. Specifically, G2MILP\nrepresents MILP instances as bipartite graphs, and applies a masked variational\nautoencoder to iteratively corrupt and replace parts of the original graphs to\ngenerate new ones. The appealing feature of G2MILP is that it can learn to\ngenerate novel and realistic MILP instances without prior expert-designed\nformulations, while preserving the structures and computational hardness of\nreal-world datasets, simultaneously. Thus the generated instances can\nfacilitate downstream tasks for enhancing MILP solvers under limited data\navailability. We design a suite of benchmarks to evaluate the quality of the\ngenerated MILP instances. Experiments demonstrate that our method can produce\ninstances that closely resemble real-world datasets in terms of both structures\nand computational hardness. The deliverables are released at\nhttps://miralab-ustc.github.io/L2O-G2MILP.\n", "date": "2023-10-04", "categories": []} {"id": "2403.04264", "title": "Competitive Facility Location under Random Utilities and Routing\n Constraints", "abstract": " In this paper, we study a facility location problem within a competitive\nmarket context, where customer demand is predicted by a random utility choice\nmodel. Unlike prior research, which primarily focuses on simple constraints\nsuch as a cardinality constraint on the number of selected locations, we\nintroduce routing constraints that necessitate the selection of locations in a\nmanner that guarantees the existence of a tour visiting all chosen locations\nwhile adhering to a specified tour length upper bound. Such routing constraints\nfind crucial applications in various real-world scenarios. The problem at hand\nfeatures a non-linear objective function, resulting from the utilization of\nrandom utilities, together with complex routing constraints, making it\ncomputationally challenging. To tackle this problem, we explore three types of\nvalid cuts, namely, outer-approximation and submodular cuts to handle the\nnonlinear objective function, as well as sub-tour elimination cuts to address\nthe complex routing constraints. These lead to the development of two exact\nsolution methods: a nested cutting plane and nested branch-and-cut algorithms,\nwhere these valid cuts are iteratively added to a master problem through two\nnested loops. We also prove that our nested cutting plane method always\nconverges to optimality after a finite number of iterations. Furthermore, we\ndevelop a local search-based metaheuristic tailored for solving large-scale\ninstances and show its pros and cons compared to exact methods. Extensive\nexperiments are conducted on problem instances of varying sizes, demonstrating\nthat our approach excels in terms of solution quality and computation time when\ncompared to other baseline approaches.\n", "date": "2024-03-07", "categories": []} {"id": "2403.06466", "title": "RL-MSA: a Reinforcement Learning-based Multi-line bus Scheduling\n Approach", "abstract": " Multiple Line Bus Scheduling Problem (MLBSP) is vital to save operational\ncost of bus company and guarantee service quality for passengers. Existing\napproaches typically generate a bus scheduling scheme in an offline manner and\nthen schedule buses according to the scheme. In practice, uncertain events such\nas traffic congestion occur frequently, which may make the pre-determined bus\nscheduling scheme infeasible. In this paper, MLBSP is modeled as a Markov\nDecision Process (MDP). A Reinforcement Learning-based Multi-line bus\nScheduling Approach (RL-MSA) is proposed for bus scheduling at both the offline\nand online phases. At the offline phase, deadhead decision is integrated into\nbus selection decision for the first time to simplify the learning problem. At\nthe online phase, deadhead decision is made through a time window mechanism\nbased on the policy learned at the offline phase. We develop several new and\nuseful state features including the features for control points, bus lines and\nbuses. A bus priority screening mechanism is invented to construct bus-related\nfeatures. Considering the interests of both the bus company and passengers, a\nreward function combining the final reward and the step-wise reward is devised.\nExperiments at the offline phase demonstrate that the number of buses used of\nRL-MSA is decreased compared with offline optimization approaches. At the\nonline phase, RL-MSA can cover all departure times in a timetable (i.e.,\nservice quality) without increasing the number of buses used (i.e., operational\ncost).\n", "date": "2024-03-11", "categories": []} {"id": "2403.05318", "title": "Looking Ahead to Avoid Being Late: Solving Hard-Constrained Traveling\n Salesman Problem", "abstract": " Many real-world problems can be formulated as a constrained Traveling\nSalesman Problem (TSP). However, the constraints are always complex and\nnumerous, making the TSPs challenging to solve. When the number of complicated\nconstraints grows, it is time-consuming for traditional heuristic algorithms to\navoid illegitimate outcomes. Learning-based methods provide an alternative to\nsolve TSPs in a soft manner, which also supports GPU acceleration to generate\nsolutions quickly. Nevertheless, the soft manner inevitably results in\ndifficulty solving hard-constrained problems with learning algorithms, and the\nconflicts between legality and optimality may substantially affect the\noptimality of the solution. To overcome this problem and to have an effective\nsolution against hard constraints, we proposed a novel learning-based method\nthat uses looking-ahead information as the feature to improve the legality of\nTSP with Time Windows (TSPTW) solutions. Besides, we constructed TSPTW datasets\nwith hard constraints in order to accurately evaluate and benchmark the\nstatistical performance of various approaches, which can serve the community\nfor future research. With comprehensive experiments on diverse datasets, MUSLA\noutperforms existing baselines and shows generalizability potential.\n", "date": "2024-03-08", "categories": []} {"id": "2207.14744", "title": "Enhanced Methods for the Weight Constrained Shortest Path Problem", "abstract": " The classic problem of constrained pathfinding is a well-studied, yet\nchallenging, topic in AI with a broad range of applications in various areas\nsuch as communication and transportation. The Weight Constrained Shortest Path\nProblem (WCSPP), the base form of constrained pathfinding with only one side\nconstraint, aims to plan a cost-optimum path with limited weight/resource\nusage. Given the bi-criteria nature of the problem (i.e., dealing with the cost\nand weight of paths), methods addressing the WCSPP have some common properties\nwith bi-objective search. This paper leverages the recent state-of-the-art\ntechniques in both constrained pathfinding and bi-objective search and presents\ntwo new solution approaches to the WCSPP on the basis of A* search, both\ncapable of solving hard WCSPP instances on very large graphs. We empirically\nevaluate the performance of our algorithms on a set of large and realistic\nproblem instances and show their advantages over the state-of-the-art\nalgorithms in both time and space metrics. This paper also investigates the\nimportance of priority queues in constrained search with A*. We show with\nextensive experiments on both realistic and randomised graphs how bucket-based\nqueues without tie-breaking can effectively improve the algorithmic performance\nof exhaustive A*-based bi-criteria searches.\n", "date": "2022-07-29", "categories": []} {"id": "2305.01883", "title": "A Lightweight CNN-Transformer Model for Learning Traveling Salesman\n Problems", "abstract": " Several studies have attempted to solve traveling salesman problems (TSPs)\nusing various deep learning techniques. Among them, Transformer-based models\nshow state-of-the-art performance even for large-scale Traveling Salesman\nProblems (TSPs). However, they are based on fully-connected attention models\nand suffer from large computational complexity and GPU memory usage. Our work\nis the first CNN-Transformer model based on a CNN embedding layer and partial\nself-attention for TSP. Our CNN-Transformer model is able to better learn\nspatial features from input data using a CNN embedding layer compared with the\nstandard Transformer-based models. It also removes considerable redundancy in\nfully-connected attention models using the proposed partial self-attention.\nExperimental results show that the proposed CNN embedding layer and partial\nself-attention are very effective in improving performance and computational\ncomplexity. The proposed model exhibits the best performance in real-world\ndatasets and outperforms other existing state-of-the-art (SOTA)\nTransformer-based models in various aspects. Our code is publicly available at\nhttps://github.com/cm8908/CNN_Transformer3.\n", "date": "2023-05-03", "categories": []} {"id": "2403.03517", "title": "IB-Net: Initial Branch Network for Variable Decision in Boolean\n Satisfiability", "abstract": " Boolean Satisfiability problems are vital components in Electronic Design\nAutomation, particularly within the Logic Equivalence Checking process.\nCurrently, SAT solvers are employed for these problems and neural network is\ntried as assistance to solvers. However, as SAT problems in the LEC context are\ndistinctive due to their predominantly unsatisfiability nature and a\nsubstantial proportion of UNSAT-core variables, existing neural network\nassistance has proven unsuccessful in this specialized domain. To tackle this\nchallenge, we propose IB-Net, an innovative framework utilizing graph neural\nnetworks and novel graph encoding techniques to model unsatisfiable problems\nand interact with state-of-the-art solvers. Extensive evaluations across\nsolvers and datasets demonstrate IB-Net's acceleration, achieving an average\nruntime speedup of 5.0% on industrial data and 8.3% on SAT competition data\nempirically. This breakthrough advances efficient solving in LEC workflows.\n", "date": "2024-03-06", "categories": []} {"id": "2403.02783", "title": "Where the Really Hard Quadratic Assignment Problems Are: the QAP-SAT\n instances", "abstract": " The Quadratic Assignment Problem (QAP) is one of the major domains in the\nfield of evolutionary computation, and more widely in combinatorial\noptimization. This paper studies the phase transition of the QAP, which can be\ndescribed as a dramatic change in the problem's computational complexity and\nsatisfiability, within a narrow range of the problem parameters. To approach\nthis phenomenon, we introduce a new QAP-SAT design of the initial problem based\non submodularity to capture its difficulty with new features. This\ndecomposition is studied experimentally using branch-and-bound and tabu search\nsolvers. A phase transition parameter is then proposed. The critical parameter\nof phase transition satisfaction and that of the solving effort are shown to be\nhighly correlated for tabu search, thus allowing the prediction of difficult\ninstances.\n", "date": "2024-03-05", "categories": []} {"id": "2310.06543", "title": "An Edge-Aware Graph Autoencoder Trained on Scale-Imbalanced Data for\n Traveling Salesman Problems", "abstract": " In recent years, there has been a notable surge in research on machine\nlearning techniques for combinatorial optimization. It has been shown that\nlearning-based methods outperform traditional heuristics and mathematical\nsolvers on the Traveling Salesman Problem (TSP) in terms of both performance\nand computational efficiency. However, most learning-based TSP solvers are\nprimarily designed for fixed-scale TSP instances, and also require a large\nnumber of training samples to achieve optimal performance. To fill this gap,\nthis work proposes a data-driven graph representation learning method for\nsolving TSPs with various numbers of cities. Specifically, we formulate the TSP\nas a link prediction task and propose an edge-aware graph autoencoder (EdgeGAE)\nmodel that can solve TSPs by learning from various-scale samples with an\nimbalanced distribution. A residual gated encoder is trained to learn latent\nedge embeddings, followed by an edge-centered decoder to output link\npredictions in an end-to-end manner. Furthermore, we introduce an active\nsampling strategy into the training process to improve the model's\ngeneralization capability in large-scale scenarios. To investigate the model's\npractical applicability, we generate a scale-imbalanced dataset comprising\n50,000 TSP instances ranging from 50 to 500 cities. The experimental results\ndemonstrate that the proposed edge-aware graph autoencoder model achieves a\nhighly competitive performance among state-of-the-art graph learning-based\napproaches in solving TSPs with various scales, implying its remarkable\npotential in dealing with practical optimization challenges.\n", "date": "2023-10-10", "categories": []} {"id": "2310.15516", "title": "Graph Attention-based Deep Reinforcement Learning for solving the\n Chinese Postman Problem with Load-dependent costs", "abstract": " Recently, Deep reinforcement learning (DRL) models have shown promising\nresults in solving routing problems. However, most DRL solvers are commonly\nproposed to solve node routing problems, such as the Traveling Salesman Problem\n(TSP). Meanwhile, there has been limited research on applying neural methods to\narc routing problems, such as the Chinese Postman Problem (CPP), since they\noften feature irregular and complex solution spaces compared to TSP. To fill\nthese gaps, this paper proposes a novel DRL framework to address the CPP with\nload-dependent costs (CPP-LC) (Corberan et al., 2018), which is a complex arc\nrouting problem with load constraints. The novelty of our method is two-fold.\nFirst, we formulate the CPP-LC as a Markov Decision Process (MDP) sequential\nmodel. Subsequently, we introduce an autoregressive model based on DRL, namely\nArc-DRL, consisting of an encoder and decoder to address the CPP-LC challenge\neffectively. Such a framework allows the DRL model to work efficiently and\nscalably to arc routing problems. Furthermore, we propose a new bio-inspired\nmeta-heuristic solution based on Evolutionary Algorithm (EA) for CPP-LC.\nExtensive experiments show that Arc-DRL outperforms existing meta-heuristic\nmethods such as Iterative Local Search (ILS) and Variable Neighborhood Search\n(VNS) proposed by (Corberan et al., 2018) on large benchmark datasets for\nCPP-LC regarding both solution quality and running time; while the EA gives the\nbest solution quality with much more running time. We release our C++\nimplementations for metaheuristics such as EA, ILS and VNS along with the code\nfor data generation and our generated data at\nhttps://github.com/HySonLab/Chinese_Postman_Problem\n", "date": "2023-10-24", "categories": []} {"id": "2403.00859", "title": "Team Formation amidst Conflicts", "abstract": " In this work, we formulate the problem of team formation amidst conflicts.\nThe goal is to assign individuals to tasks, with given capacities, taking into\naccount individuals' task preferences and the conflicts between them. Using\ndependent rounding schemes as our main toolbox, we provide efficient\napproximation algorithms. Our framework is extremely versatile and can model\nmany different real-world scenarios as they arise in educational settings and\nhuman-resource management. We test and deploy our algorithms on real-world\ndatasets and we show that our algorithms find assignments that are better than\nthose found by natural baselines. In the educational setting we also show how\nour assignments are far better than those done manually by human experts. In\nthe human resource management application we show how our assignments increase\nthe diversity of teams. Finally, using a synthetic dataset we demonstrate that\nour algorithms scale very well in practice.\n", "date": "2024-02-29", "categories": []} {"id": "2305.00188", "title": "New Characterizations and Efficient Local Search for General Integer\n Linear Programming", "abstract": " Integer linear programming (ILP) models a wide range of practical\ncombinatorial optimization problems and significantly impacts industry and\nmanagement sectors. This work proposes new characterizations of ILP with the\nconcept of boundary solutions. Motivated by the new characterizations, we\ndevelop a new local search algorithm Local-ILP, which is efficient for solving\ngeneral ILP validated on a large heterogeneous problem dataset. We propose a\nnew local search framework that switches between three modes, namely Search,\nImprove, and Restore modes. Two new operators are proposed, namely the tight\nmove and the lift move operators, which are associated with appropriate scoring\nfunctions. Different modes apply different operators to realize different\nsearch strategies and the algorithm switches between three modes according to\nthe current search state. Putting these together, we develop a local search ILP\nsolver called Local-ILP. Experiments conducted on the MIPLIB dataset show the\neffectiveness of our algorithm in solving large-scale hard ILP problems. In the\naspect of finding a good feasible solution quickly, Local-ILP is competitive\nand complementary to the state-of-the-art commercial solver Gurobi and\nsignificantly outperforms the state-of-the-art non-commercial solver SCIP.\nMoreover, our algorithm establishes new records for 6 MIPLIB open instances.\nThe theoretical analysis of our algorithm is also presented, which shows our\nalgorithm could avoid visiting unnecessary regions.\n", "date": "2023-04-29", "categories": []} {"id": "2403.00016", "title": "Deep Sensitivity Analysis for Objective-Oriented Combinatorial\n Optimization", "abstract": " Pathogen control is a critical aspect of modern poultry farming, providing\nimportant benefits for both public health and productivity. Effective poultry\nmanagement measures to reduce pathogen levels in poultry flocks promote food\nsafety by lowering risks of food-borne illnesses. They also support animal\nhealth and welfare by preventing infectious diseases that can rapidly spread\nand impact flock growth, egg production, and overall health. This study frames\nthe search for optimal management practices that minimize the presence of\nmultiple pathogens as a combinatorial optimization problem. Specifically, we\nmodel the various possible combinations of management settings as a solution\nspace that can be efficiently explored to identify configurations that\noptimally reduce pathogen levels. This design incorporates a neural network\nfeedback-based method that combines feature explanations with global\nsensitivity analysis to ensure combinatorial optimization in multiobjective\nsettings. Our preliminary experiments have promising results when applied to\ntwo real-world agricultural datasets. While further validation is still needed,\nthese early experimental findings demonstrate the potential of the model to\nderive targeted feature interactions that adaptively optimize pathogen control\nunder varying real-world constraints.\n", "date": "2024-02-28", "categories": []} {"id": "2403.00026", "title": "Learning to Deliver: a Foundation Model for the Montreal Capacitated\n Vehicle Routing Problem", "abstract": " In this paper, we present the Foundation Model for the Montreal Capacitated\nVehicle Routing Problem (FM-MCVRP), a novel Deep Learning (DL) model that\napproximates high-quality solutions to a variant of the Capacitated Vehicle\nRouting Problem (CVRP) that characterizes many real-world applications. The\nso-called Montreal Capacitated Vehicle Routing Problem (MCVRP), first formally\ndescribed by Bengio et al. (2021), is defined on a fixed and finite graph,\nwhich is analogous to a city. Each MCVRP instance is essentially the sub-graph\nconnecting a randomly sampled subset of the nodes in the fixed graph, which\nrepresent a set of potential addresses in a real-world delivery problem on a\ngiven day. Our work exploits this problem structure to frame the MCVRP as an\nanalogous Natural Language Processing (NLP) task. Specifically, we leverage a\nTransformer architecture embedded in a Large Language Model (LLM) framework to\ntrain our model in a supervised manner on computationally inexpensive,\nsub-optimal MCVRP solutions obtained algorithmically. Through comprehensive\ncomputational experiments, we show that FM-MCVRP produces better MCVRP\nsolutions than the training data and generalizes to larger sized problem\ninstances not seen during training. Even when compared to near-optimal\nsolutions from state-of-the-art heuristics, FM-MCVRP yields competitive results\ndespite being trained on inferior data. For instance, for 400-customer\nproblems, FM-MCVRP solutions on average fall within 2% of the benchmark. Our\nresults further demonstrate that unlike prior works in the literature, FM-MCVRP\nis a unified model, which performs consistently and reliably on a range of\nproblem instance sizes and parameter values such as the vehicle capacity.\n", "date": "2024-02-28", "categories": []} {"id": "2312.10797", "title": "Large-Scale Multi-Robot Coverage Path Planning via Local Search", "abstract": " We study graph-based Multi-Robot Coverage Path Planning (MCPP) that aims to\ncompute coverage paths for multiple robots to cover all vertices of a given 2D\ngrid terrain graph $G$. Existing graph-based MCPP algorithms first compute a\ntree cover on $G$ -- a forest of multiple trees that cover all vertices -- and\nthen employ the Spanning Tree Coverage (STC) paradigm to generate coverage\npaths on the decomposed graph $D$ of the terrain graph $G$ by circumnavigating\nthe edges of the computed trees, aiming to optimize the makespan (i.e., the\nmaximum coverage path cost among all robots). In this paper, we take a\ndifferent approach by exploring how to systematically search for good coverage\npaths directly on $D$. We introduce a new algorithmic framework, called\nLS-MCPP, which leverages a local search to operate directly on $D$. We propose\na novel standalone paradigm, Extended-STC (ESTC), that extends STC to achieve\ncomplete coverage for MCPP on any decomposed graphs, even those resulting from\nincomplete terrain graphs. Furthermore, we demonstrate how to integrate ESTC\nwith three novel types of neighborhood operators into our framework to\neffectively guide its search process. Our extensive experiments demonstrate the\neffectiveness of LS-MCPP, consistently improving the initial solution returned\nby two state-of-the-art baseline algorithms that compute suboptimal tree covers\non $G$, with a notable reduction in makespan by up to 35.7\\% and 30.3\\%,\nrespectively. Moreover, LS-MCPP consistently matches or surpasses the results\nof optimal tree cover computation, achieving these outcomes with orders of\nmagnitude faster runtime, thereby showcasing its significant benefits for\nlarge-scale real-world coverage tasks.\n", "date": "2023-12-17", "categories": []} {"id": "2307.08816", "title": "Accelerating Cutting-Plane Algorithms via Reinforcement Learning\n Surrogates", "abstract": " Discrete optimization belongs to the set of $\\mathcal{NP}$-hard problems,\nspanning fields such as mixed-integer programming and combinatorial\noptimization. A current standard approach to solving convex discrete\noptimization problems is the use of cutting-plane algorithms, which reach\noptimal solutions by iteratively adding inequalities known as \\textit{cuts} to\nrefine a feasible set. Despite the existence of a number of general-purpose\ncut-generating algorithms, large-scale discrete optimization problems continue\nto suffer from intractability. In this work, we propose a method for\naccelerating cutting-plane algorithms via reinforcement learning. Our approach\nuses learned policies as surrogates for $\\mathcal{NP}$-hard elements of the cut\ngenerating procedure in a way that (i) accelerates convergence, and (ii)\nretains guarantees of optimality. We apply our method on two types of problems\nwhere cutting-plane algorithms are commonly used: stochastic optimization, and\nmixed-integer quadratic programming. We observe the benefits of our method when\napplied to Benders decomposition (stochastic optimization) and iterative loss\napproximation (quadratic programming), achieving up to $45\\%$ faster average\nconvergence when compared to modern alternative algorithms.\n", "date": "2023-07-17", "categories": []} {"id": "2402.14847", "title": "Deep learning-driven scheduling algorithm for a single machine problem\n minimizing the total tardiness", "abstract": " In this paper, we investigate the use of the deep learning method for solving\na well-known NP-hard single machine scheduling problem with the objective of\nminimizing the total tardiness. We propose a deep neural network that acts as a\npolynomial-time estimator of the criterion value used in a single-pass\nscheduling algorithm based on Lawler's decomposition and symmetric\ndecomposition proposed by Della Croce et al. Essentially, the neural network\nguides the algorithm by estimating the best splitting of the problem into\nsubproblems. The paper also describes a new method for generating the training\ndata set, which speeds up the training dataset generation and reduces the\naverage optimality gap of solutions. The experimental results show that our\nmachine learning-driven approach can efficiently generalize information from\nthe training phase to significantly larger instances. Even though the instances\nused in the training phase have from 75 to 100 jobs, the average optimality gap\non instances with up to 800 jobs is 0.26%, which is almost five times less than\nthe gap of the state-of-the-art heuristic.\n", "date": "2024-02-19", "categories": []} {"id": "2402.14600", "title": "Diffusion Model-Based Multiobjective Optimization for Gasoline Blending\n Scheduling", "abstract": " Gasoline blending scheduling uses resource allocation and operation\nsequencing to meet a refinery's production requirements. The presence of\nnonlinearity, integer constraints, and a large number of decision variables\nadds complexity to this problem, posing challenges for traditional and\nevolutionary algorithms. This paper introduces a novel multiobjective\noptimization approach driven by a diffusion model (named DMO), which is\ndesigned specifically for gasoline blending scheduling. To address integer\nconstraints and generate feasible schedules, the diffusion model creates\nmultiple intermediate distributions between Gaussian noise and the feasible\ndomain. Through iterative processes, the solutions transition from Gaussian\nnoise to feasible schedules while optimizing the objectives using the gradient\ndescent method. DMO achieves simultaneous objective optimization and constraint\nadherence. Comparative tests are conducted to evaluate DMO's performance across\nvarious scales. The experimental results demonstrate that DMO surpasses\nstate-of-the-art multiobjective evolutionary algorithms in terms of efficiency\nwhen solving gasoline blending scheduling problems.\n", "date": "2024-02-04", "categories": []} {"id": "2308.02787", "title": "Solving Logistic-Oriented Bin Packing Problems Through a Hybrid\n Quantum-Classical Approach", "abstract": " The Bin Packing Problem is a classic problem with wide industrial\napplicability. In fact, the efficient packing of items into bins is one of the\ntoughest challenges in many logistic corporations and is a critical issue for\nreducing storage costs or improving vehicle space allocation. In this work, we\nresort to our previously published quantum-classical framework known as\nQ4RealBPP, and elaborate on the solving of real-world oriented instances of the\nBin Packing Problem. With this purpose, this paper gravitates on the following\ncharacteristics: i) the existence of heterogeneous bins, ii) the extension of\nthe framework to solve not only three-dimensional, but also one- and\ntwo-dimensional instances of the problem, iii) requirements for item-bin\nassociations, and iv) delivery priorities. All these features have been tested\nin this paper, as well as the ability of Q4RealBPP to solve real-world oriented\ninstances.\n", "date": "2023-08-05", "categories": []} {"id": "2402.09765", "title": "Reinforcement Learning for Solving Stochastic Vehicle Routing Problem\n with Time Windows", "abstract": " This paper introduces a reinforcement learning approach to optimize the\nStochastic Vehicle Routing Problem with Time Windows (SVRP), focusing on\nreducing travel costs in goods delivery. We develop a novel SVRP formulation\nthat accounts for uncertain travel costs and demands, alongside specific\ncustomer time windows. An attention-based neural network trained through\nreinforcement learning is employed to minimize routing costs. Our approach\naddresses a gap in SVRP research, which traditionally relies on heuristic\nmethods, by leveraging machine learning. The model outperforms the Ant-Colony\nOptimization algorithm, achieving a 1.73% reduction in travel costs. It\nuniquely integrates external information, demonstrating robustness in diverse\nenvironments, making it a valuable benchmark for future SVRP studies and\nindustry application.\n", "date": "2024-02-15", "categories": []} {"id": "2402.09961", "title": "Enhancing Courier Scheduling in Crowdsourced Last-Mile Delivery through\n Dynamic Shift Extensions: A Deep Reinforcement Learning Approach", "abstract": " Crowdsourced delivery platforms face complex scheduling challenges to match\ncouriers and customer orders. We consider two types of crowdsourced couriers,\nnamely, committed and occasional couriers, each with different compensation\nschemes. Crowdsourced delivery platforms usually schedule committed courier\nshifts based on predicted demand. Therefore, platforms may devise an offline\nschedule for committed couriers before the planning period. However, due to the\nunpredictability of demand, there are instances where it becomes necessary to\nmake online adjustments to the offline schedule. In this study, we focus on the\nproblem of dynamically adjusting the offline schedule through shift extensions\nfor committed couriers. This problem is modeled as a sequential decision\nprocess. The objective is to maximize platform profit by determining the shift\nextensions of couriers and the assignments of requests to couriers. To solve\nthe model, a Deep Q-Network (DQN) learning approach is developed. Comparing\nthis model with the baseline policy where no extensions are allowed\ndemonstrates the benefits that platforms can gain from allowing shift\nextensions in terms of reward, reduced lost order costs, and lost requests.\nAdditionally, sensitivity analysis showed that the total extension compensation\nincreases in a nonlinear manner with the arrival rate of requests, and in a\nlinear manner with the arrival rate of occasional couriers. On the compensation\nsensitivity, the results showed that the normal scenario exhibited the highest\naverage number of shift extensions and, consequently, the fewest average number\nof lost requests. These findings serve as evidence of the successful learning\nof such dynamics by the DQN algorithm.\n", "date": "2024-02-15", "categories": []} {"id": "2402.10177", "title": "Large Scale Constrained Clustering With Reinforcement Learning", "abstract": " Given a network, allocating resources at clusters level, rather than at each\nnode, enhances efficiency in resource allocation and usage. In this paper, we\nstudy the problem of finding fully connected disjoint clusters to minimize the\nintra-cluster distances and maximize the number of nodes assigned to the\nclusters, while also ensuring that no two nodes within a cluster exceed a\nthreshold distance. While the problem can easily be formulated using a binary\nlinear model, traditional combinatorial optimization solvers struggle when\ndealing with large-scale instances. We propose an approach to solve this\nconstrained clustering problem via reinforcement learning. Our method involves\ntraining an agent to generate both feasible and (near) optimal solutions. The\nagent learns problem-specific heuristics, tailored to the instances encountered\nin this task. In the results section, we show that our algorithm finds near\noptimal solutions, even for large scale instances.\n", "date": "2024-02-15", "categories": []} {"id": "2211.10936", "title": "Deep Reinforcement Learning Guided Improvement Heuristic for Job Shop\n Scheduling", "abstract": " Recent studies in using deep reinforcement learning (DRL) to solve Job-shop\nscheduling problems (JSSP) focus on construction heuristics. However, their\nperformance is still far from optimality, mainly because the underlying graph\nrepresentation scheme is unsuitable for modelling partial solutions at each\nconstruction step. This paper proposes a novel DRL-guided improvement heuristic\nfor solving JSSP, where graph representation is employed to encode complete\nsolutions. We design a Graph Neural-Network-based representation scheme,\nconsisting of two modules to effectively capture the information of dynamic\ntopology and different types of nodes in graphs encountered during the\nimprovement process. To speed up solution evaluation during improvement, we\npresent a novel message-passing mechanism that can evaluate multiple solutions\nsimultaneously. We prove that the computational complexity of our method scales\nlinearly with problem size. Experiments on classic benchmarks show that the\nimprovement policy learned by our method outperforms state-of-the-art DRL-based\nmethods by a large margin.\n", "date": "2022-11-20", "categories": []} {"id": "2310.09196", "title": "A 4-approximation algorithm for min max correlation clustering", "abstract": " We introduce a lower bounding technique for the min max correlation\nclustering problem and, based on this technique, a combinatorial\n4-approximation algorithm for complete graphs. This improves upon the previous\nbest known approximation guarantees of 5, using a linear program formulation\n(Kalhan et al., 2019), and 40, for a combinatorial algorithm (Davies et al.,\n2023a). We extend this algorithm by a greedy joining heuristic and show\nempirically that it improves the state of the art in solution quality and\nruntime on several benchmark datasets.\n", "date": "2023-10-13", "categories": []} {"id": "2402.08979", "title": "Learning-enabled Flexible Job-shop Scheduling for Scalable Smart\n Manufacturing", "abstract": " In smart manufacturing systems (SMSs), flexible job-shop scheduling with\ntransportation constraints (FJSPT) is essential to optimize solutions for\nmaximizing productivity, considering production flexibility based on automated\nguided vehicles (AGVs). Recent developments in deep reinforcement learning\n(DRL)-based methods for FJSPT have encountered a scale generalization\nchallenge. These methods underperform when applied to environment at scales\ndifferent from their training set, resulting in low-quality solutions. To\naddress this, we introduce a novel graph-based DRL method, named the\nHeterogeneous Graph Scheduler (HGS). Our method leverages locally extracted\nrelational knowledge among operations, machines, and vehicle nodes for\nscheduling, with a graph-structured decision-making framework that reduces\nencoding complexity and enhances scale generalization. Our performance\nevaluation, conducted with benchmark datasets, reveals that the proposed method\noutperforms traditional dispatching rules, meta-heuristics, and existing\nDRL-based approaches in terms of makespan performance, even on large-scale\ninstances that have not been experienced during training.\n", "date": "2024-02-14", "categories": []} {"id": "2306.06064", "title": "Neural Algorithmic Reasoning for Combinatorial Optimisation", "abstract": " Solving NP-hard/complete combinatorial problems with neural networks is a\nchallenging research area that aims to surpass classical approximate\nalgorithms. The long-term objective is to outperform hand-designed heuristics\nfor NP-hard/complete problems by learning to generate superior solutions solely\nfrom training data. Current neural-based methods for solving CO problems often\noverlook the inherent \"algorithmic\" nature of the problems. In contrast,\nheuristics designed for CO problems, e.g. TSP, frequently leverage\nwell-established algorithms, such as those for finding the minimum spanning\ntree. In this paper, we propose leveraging recent advancements in neural\nalgorithmic reasoning to improve the learning of CO problems. Specifically, we\nsuggest pre-training our neural model on relevant algorithms before training it\non CO instances. Our results demonstrate that by using this learning setup, we\nachieve superior performance compared to non-algorithmically informed deep\nlearning models.\n", "date": "2023-05-18", "categories": []} {"id": "2402.08246", "title": "Ant Colony Optimization for Cooperative Inspection Path Planning Using\n Multiple Unmanned Aerial Vehicles", "abstract": " This paper presents a new swarm intelligence-based approach to deal with the\ncooperative path planning problem of unmanned aerial vehicles (UAVs), which is\nessential for the automatic inspection of infrastructure. The approach uses a\n3D model of the structure to generate viewpoints for the UAVs. The calculation\nof the viewpoints considers the constraints related to the UAV formation model,\ncamera parameters, and requirements for data post-processing. The viewpoints\nare then used as input to formulate the path planning as an extended traveling\nsalesman problem and the definition of a new cost function. Ant colony\noptimization is finally used to solve the problem to yield optimal inspection\npaths. Experiments with 3D models of real structures have been conducted to\nevaluate the performance of the proposed approach. The results show that our\nsystem is not only capable of generating feasible inspection paths for UAVs but\nalso reducing the path length by 29.47\\% for complex structures when compared\nwith another heuristic approach. The source code of the algorithm can be found\nat https://github.com/duynamrcv/aco_3d_ipp.\n", "date": "2024-02-13", "categories": []} {"id": "2402.06968", "title": "Contextual Stochastic Vehicle Routing with Time Windows", "abstract": " We study the vehicle routing problem with time windows (VRPTW) and stochastic\ntravel times, in which the decision-maker observes related contextual\ninformation, represented as feature variables, before making routing decisions.\nDespite the extensive literature on stochastic VRPs, the integration of feature\nvariables has received limited attention in this context. We introduce the\ncontextual stochastic VRPTW, which minimizes the total transportation cost and\nexpected late arrival penalties conditioned on the observed features. Since the\njoint distribution of travel times and features is unknown, we present novel\ndata-driven prescriptive models that use historical data to provide an\napproximate solution to the problem. We distinguish the prescriptive models\nbetween point-based approximation, sample average approximation, and\npenalty-based approximation, each taking a different perspective on dealing\nwith stochastic travel times and features. We develop specialized\nbranch-price-and-cut algorithms to solve these data-driven prescriptive models.\nIn our computational experiments, we compare the out-of-sample cost performance\nof different methods on instances with up to one hundred customers. Our results\nshow that, surprisingly, a feature-dependent sample average approximation\noutperforms existing and novel methods in most settings.\n", "date": "2024-02-10", "categories": []} {"id": "2402.04907", "title": "On a Combinatorial Problem Arising in Machine Teaching", "abstract": " We study a model of machine teaching where the teacher mapping is constructed\nfrom a size function on both concepts and examples. The main question in\nmachine teaching is the minimum number of examples needed for any concept, the\nso-called teaching dimension. A recent paper [7] conjectured that the worst\ncase for this model, as a function of the size of the concept class, occurs\nwhen the consistency matrix contains the binary representations of numbers from\nzero and up. In this paper we prove their conjecture. The result can be seen as\na generalization of a theorem resolving the edge isoperimetry problem for\nhypercubes [12], and our proof is based on a lemma of [10].\n", "date": "2024-02-07", "categories": []} {"id": "2402.05501", "title": "Machine Learning Augmented Branch and Bound for Mixed Integer Linear\n Programming", "abstract": " Mixed Integer Linear Programming (MILP) is a pillar of mathematical\noptimization that offers a powerful modeling language for a wide range of\napplications. During the past decades, enormous algorithmic progress has been\nmade in solving MILPs, and many commercial and academic software packages\nexist. Nevertheless, the availability of data, both from problem instances and\nfrom solvers, and the desire to solve new problems and larger (real-life)\ninstances, trigger the need for continuing algorithmic development. MILP\nsolvers use branch and bound as their main component. In recent years, there\nhas been an explosive development in the use of machine learning algorithms for\nenhancing all main tasks involved in the branch-and-bound algorithm, such as\nprimal heuristics, branching, cutting planes, node selection and solver\nconfiguration decisions. This paper presents a survey of such approaches,\naddressing the vision of integration of machine learning and mathematical\noptimization as complementary technologies, and how this integration can\nbenefit MILP solving. In particular, we give detailed attention to machine\nlearning algorithms that automatically optimize some metric of branch-and-bound\nefficiency. We also address how to represent MILPs in the context of applying\nlearning algorithms, MILP benchmarks and software.\n", "date": "2024-02-08", "categories": []} {"id": "2402.04597", "title": "CMSA algorithm for solving the prioritized pairwise test data generation\n problem in software product lines", "abstract": " In Software Product Lines (SPLs) it may be difficult or even impossible to\ntest all the products of the family because of the large number of valid\nfeature combinations that may exist. Thus, we want to find a minimal subset of\nthe product family that allows us to test all these possible combinations\n(pairwise). Furthermore, when testing a single product is a great effort, it is\ndesirable to first test products composed of a set of priority features. This\nproblem is called Prioritized Pairwise Test Data Generation Problem.\n State-of-the-art algorithms based on Integer Linear Programming for this\nproblema are faster enough for small and medium instances. However, there\nexists some real instances that are too large to be computed with these\nalgorithms in a reasonable time because of the exponential growth of the number\nof candidate solutions. Also, these heuristics not always lead us to the best\nsolutions. In this work we propose a new approach based on a hybrid\nmetaheuristic algorithm called Construct, Merge, Solve & Adapt. We compare this\nmatheuristic with four algorithms: a Hybrid algorithm based on Integer Linear\nProgramming ((HILP), a Hybrid algorithm based on Integer Nonlinear Programming\n(HINLP), the Parallel Prioritized Genetic Solver (PPGS), and a greedy algorithm\ncalled prioritized-ICPL. The analysis reveals that CMSA results in\nstatistically significantly better quality solutions in most instances and for\nmost levels of weighted coverage, although it requires more execution time.\n", "date": "2024-02-07", "categories": []} {"id": "2207.06294", "title": "Reinforcement Learning Assisted Recursive QAOA", "abstract": " Variational quantum algorithms such as the Quantum Approximation Optimization\nAlgorithm (QAOA) in recent years have gained popularity as they provide the\nhope of using NISQ devices to tackle hard combinatorial optimization problems.\nIt is, however, known that at low depth, certain locality constraints of QAOA\nlimit its performance. To go beyond these limitations, a non-local variant of\nQAOA, namely recursive QAOA (RQAOA), was proposed to improve the quality of\napproximate solutions. The RQAOA has been studied comparatively less than QAOA,\nand it is less understood, for instance, for what family of instances it may\nfail to provide high quality solutions. However, as we are tackling\n$\\mathsf{NP}$-hard problems (specifically, the Ising spin model), it is\nexpected that RQAOA does fail, raising the question of designing even better\nquantum algorithms for combinatorial optimization. In this spirit, we identify\nand analyze cases where RQAOA fails and, based on this, propose a reinforcement\nlearning enhanced RQAOA variant (RL-RQAOA) that improves upon RQAOA. We show\nthat the performance of RL-RQAOA improves over RQAOA: RL-RQAOA is strictly\nbetter on these identified instances where RQAOA underperforms, and is\nsimilarly performing on instances where RQAOA is near-optimal. Our work\nexemplifies the potentially beneficial synergy between reinforcement learning\nand quantum (inspired) optimization in the design of new, even better\nheuristics for hard problems.\n", "date": "2022-07-13", "categories": []} {"id": "2402.01849", "title": "Capturing waste collection planning expert knowledge in a fitness\n function through preference learning", "abstract": " This paper copes with the COGERSA waste collection process. Up to now,\nexperts have been manually designed the process using a trial and error\nmechanism. This process is not globally optimized, since it has been\nprogressively and locally built as council demands appear. Planning\noptimization algorithms usually solve it, but they need a fitness function to\nevaluate a route planning quality. The drawback is that even experts are not\nable to propose one in a straightforward way due to the complexity of the\nprocess. Hence, the goal of this paper is to build a fitness function though a\npreference framework, taking advantage of the available expert knowledge and\nexpertise. Several key performance indicators together with preference\njudgments are carefully established according to the experts for learning a\npromising fitness function. Particularly, the additivity property of them makes\nthe task be much more affordable, since it allows to work with routes rather\nthan with route plannings. Besides, a feature selection analysis is performed\nover such indicators, since the experts suspect of a potential existing (but\nunknown) redundancy among them. The experiment results confirm this hypothesis,\nsince the best $C-$index ($98\\%$ against around $94\\%$) is reached when 6 or 8\nout of 21 indicators are taken. Particularly, truck load seems to be a highly\npromising key performance indicator, together to the travelled distance along\nnon-main roads. A comparison with other existing approaches shows that the\nproposed method clearly outperforms them, since the $C-$index goes from $72\\%$\nor $90\\%$ to $98\\%$.\n", "date": "2024-02-02", "categories": []} {"id": "2402.03525", "title": "Deep Reinforcement Learning for Picker Routing Problem in Warehousing", "abstract": " Order Picker Routing is a critical issue in Warehouse Operations Management.\nDue to the complexity of the problem and the need for quick solutions,\nsuboptimal algorithms are frequently employed in practice. However,\nReinforcement Learning offers an appealing alternative to traditional\nheuristics, potentially outperforming existing methods in terms of speed and\naccuracy. We introduce an attention based neural network for modeling picker\ntours, which is trained using Reinforcement Learning. Our method is evaluated\nagainst existing heuristics across a range of problem parameters to demonstrate\nits efficacy. A key advantage of our proposed method is its ability to offer an\noption to reduce the perceived complexity of routes.\n", "date": "2024-02-05", "categories": []} {"id": "2402.03640", "title": "torchmSAT: A GPU-Accelerated Approximation To The Maximum Satisfiability\n Problem", "abstract": " The remarkable achievements of machine learning techniques in analyzing\ndiscrete structures have drawn significant attention towards their integration\ninto combinatorial optimization algorithms. Typically, these methodologies\nimprove existing solvers by injecting learned models within the solving loop to\nenhance the efficiency of the search process. In this work, we derive a single\ndifferentiable function capable of approximating solutions for the Maximum\nSatisfiability Problem (MaxSAT). Then, we present a novel neural network\narchitecture to model our differentiable function, and progressively solve\nMaxSAT using backpropagation. This approach eliminates the need for labeled\ndata or a neural network training phase, as the training process functions as\nthe solving algorithm. Additionally, we leverage the computational power of\nGPUs to accelerate these computations. Experimental results on challenging\nMaxSAT instances show that our proposed methodology outperforms two existing\nMaxSAT solvers, and is on par with another in terms of solution cost, without\nnecessitating any training or access to an underlying SAT solver. Given that\nnumerous NP-hard problems can be reduced to MaxSAT, our novel technique paves\nthe way for a new generation of solvers poised to benefit from neural network\nGPU acceleration.\n", "date": "2024-02-06", "categories": []} {"id": "2402.03736", "title": "An Effective Branch-and-Bound Algorithm with New Bounding Methods for\n the Maximum $s$-Bundle Problem", "abstract": " The Maximum s-Bundle Problem (MBP) addresses the task of identifying a\nmaximum s-bundle in a given graph. A graph G=(V, E) is called an s-bundle if\nits vertex connectivity is at least |V|-s, where the vertex connectivity equals\nthe minimum number of vertices whose deletion yields a disconnected or trivial\ngraph. MBP is NP-hard and holds relevance in numerous realworld scenarios\nemphasizing the vertex connectivity. Exact algorithms for MBP mainly follow the\nbranch-and-bound (BnB) framework, whose performance heavily depends on the\nquality of the upper bound on the cardinality of a maximum s-bundle and the\ninitial lower bound with graph reduction. In this work, we introduce a novel\nPartition-based Upper Bound (PUB) that leverages the graph partitioning\ntechnique to achieve a tighter upper bound compared to existing ones. To\nincrease the lower bound, we propose to do short random walks on a clique to\ngenerate larger initial solutions. Then, we propose a new BnB algorithm that\nuses the initial lower bound and PUB in preprocessing for graph reduction, and\nuses PUB in the BnB search process for branch pruning. Extensive experiments\nwith diverse s values demonstrate the significant progress of our algorithm\nover state-of-the-art BnB MBP algorithms. Moreover, our initial lower bound can\nalso be generalized to other relaxation clique problems.\n", "date": "2024-02-06", "categories": []} {"id": "2306.02689", "title": "Equity-Transformer: Solving NP-hard Min-Max Routing Problems as\n Sequential Generation with Equity Context", "abstract": " Min-max routing problems aim to minimize the maximum tour length among\nmultiple agents by having agents conduct tasks in a cooperative manner. These\nproblems include impactful real-world applications but are known as NP-hard.\nExisting methods are facing challenges, particularly in large-scale problems\nthat require the coordination of numerous agents to cover thousands of cities.\nThis paper proposes Equity-Transformer to solve large-scale min-max routing\nproblems. First, we employ sequential planning approach to address min-max\nrouting problems, allowing us to harness the powerful sequence generators\n(e.g., Transformer). Second, we propose key inductive biases that ensure\nequitable workload distribution among agents. The effectiveness of\nEquity-Transformer is demonstrated through its superior performance in two\nrepresentative min-max routing tasks: the min-max multi-agent traveling\nsalesman problem (min-max mTSP) and the min-max multi-agent pick-up and\ndelivery problem (min-max mPDP). Notably, our method achieves significant\nreductions of runtime, approximately 335 times, and cost values of about 53\\%\ncompared to a competitive heuristic (LKH3) in the case of 100 vehicles with\n1,000 cities of mTSP. We provide reproducible source code:\n\\url{https://github.com/kaist-silab/equity-transformer}.\n", "date": "2023-06-05", "categories": []} {"id": "2401.17527", "title": "Learning to Stop Cut Generation for Efficient Mixed-Integer Linear\n Programming", "abstract": " Cutting planes (cuts) play an important role in solving mixed-integer linear\nprograms (MILPs), as they significantly tighten the dual bounds and improve the\nsolving performance. A key problem for cuts is when to stop cuts generation,\nwhich is important for the efficiency of solving MILPs. However, many modern\nMILP solvers employ hard-coded heuristics to tackle this problem, which tends\nto neglect underlying patterns among MILPs from certain applications. To\naddress this challenge, we formulate the cuts generation stopping problem as a\nreinforcement learning problem and propose a novel hybrid graph representation\nmodel (HYGRO) to learn effective stopping strategies. An appealing feature of\nHYGRO is that it can effectively capture both the dynamic and static features\nof MILPs, enabling dynamic decision-making for the stopping strategies. To the\nbest of our knowledge, HYGRO is the first data-driven method to tackle the cuts\ngeneration stopping problem. By integrating our approach with modern solvers,\nexperiments demonstrate that HYGRO significantly improves the efficiency of\nsolving MILPs compared to competitive baselines, achieving up to 31%\nimprovement.\n", "date": "2024-01-31", "categories": []} {"id": "2402.00076", "title": "Exploitation Strategies in Conditional Markov Chain Search: A case study\n on the three-index assignment problem", "abstract": " The Conditional Markov Chain Search (CMCS) is a framework for automated\ndesign of metaheuristics for discrete combinatorial optimisation problems.\nGiven a set of algorithmic components such as hill climbers and mutations, CMCS\ndecides in which order to apply those components. The decisions are dictated by\nthe CMCS configuration that can be learnt offline. CMCS does not have an\nacceptance criterion; any moves are accepted by the framework. As a result, it\nis particularly good in exploration but is not as good at exploitation. In this\nstudy, we explore several extensions of the framework to improve its\nexploitation abilities. To perform a computational study, we applied the\nframework to the three-index assignment problem. The results of our experiments\nshowed that a two-stage CMCS is indeed superior to a single-stage CMCS.\n", "date": "2024-01-30", "categories": []} {"id": "2402.00334", "title": "Multi-agent Path Finding for Cooperative Autonomous Driving", "abstract": " Anticipating possible future deployment of connected and automated vehicles\n(CAVs), cooperative autonomous driving at intersections has been studied by\nmany works in control theory and intelligent transportation across decades.\nSimultaneously, recent parallel works in robotics have devised efficient\nalgorithms for multi-agent path finding (MAPF), though often in environments\nwith simplified kinematics. In this work, we hybridize insights and algorithms\nfrom MAPF with the structure and heuristics of optimizing the crossing order of\nCAVs at signal-free intersections. We devise an optimal and complete algorithm,\nOrder-based Search with Kinematics Arrival Time Scheduling (OBS-KATS), which\nsignificantly outperforms existing algorithms, fixed heuristics, and\nprioritized planning with KATS. The performance is maintained under different\nvehicle arrival rates, lane lengths, crossing speeds, and control horizon.\nThrough ablations and dissections, we offer insight on the contributing factors\nto OBS-KATS's performance. Our work is directly applicable to many similarly\nscaled traffic and multi-robot scenarios with directed lanes.\n", "date": "2024-02-01", "categories": []} {"id": "2310.15706", "title": "Solving the flexible job-shop scheduling problem through an enhanced\n deep reinforcement learning approach", "abstract": " In scheduling problems common in the industry and various real-world\nscenarios, responding in real-time to disruptive events is essential. Recent\nmethods propose the use of deep reinforcement learning (DRL) to learn policies\ncapable of generating solutions under this constraint. The objective of this\npaper is to introduce a new DRL method for solving the flexible job-shop\nscheduling problem, particularly for large instances. The approach is based on\nthe use of heterogeneous graph neural networks to a more informative graph\nrepresentation of the problem. This novel modeling of the problem enhances the\npolicy's ability to capture state information and improve its decision-making\ncapacity. Additionally, we introduce two novel approaches to enhance the\nperformance of the DRL approach: the first involves generating a diverse set of\nscheduling policies, while the second combines DRL with dispatching rules (DRs)\nconstraining the action space. Experimental results on two public benchmarks\nshow that our approach outperforms DRs and achieves superior results compared\nto three state-of-the-art DRL methods, particularly for large instances.\n", "date": "2023-10-24", "categories": []} {"id": "2310.18446", "title": "A Novel Skip Orthogonal List for Dynamic Optimal Transport Problem", "abstract": " Optimal transport is a fundamental topic that has attracted a great amount of\nattention from the optimization community in the past decades. In this paper,\nwe consider an interesting discrete dynamic optimal transport problem: can we\nefficiently update the optimal transport plan when the weights or the locations\nof the data points change? This problem is naturally motivated by several\napplications in machine learning. For example, we often need to compute the\noptimal transport cost between two different data sets; if some changes happen\nto a few data points, should we re-compute the high complexity cost function or\nupdate the cost by some efficient dynamic data structure? We are aware that\nseveral dynamic maximum flow algorithms have been proposed before, however, the\nresearch on dynamic minimum cost flow problem is still quite limited, to the\nbest of our knowledge. We propose a novel 2D Skip Orthogonal List together with\nsome dynamic tree techniques. Although our algorithm is based on the\nconventional simplex method, it can efficiently find the variable to pivot\nwithin expected $O(1)$ time, and complete each pivoting operation within\nexpected $O(|V|)$ time where $V$ is the set of all supply and demand nodes.\nSince dynamic modifications typically do not introduce significant changes, our\nalgorithm requires only a few simplex iterations in practice. So our algorithm\nis more efficient than re-computing the optimal transport cost that needs at\nleast one traversal over all $|E| = O(|V|^2)$ variables, where $|E|$ denotes\nthe number of edges in the network. Our experiments demonstrate that our\nalgorithm significantly outperforms existing algorithms in the dynamic\nscenarios.\n", "date": "2023-10-27", "categories": []} {"id": "2312.10572", "title": "Improved Anonymous Multi-Agent Path Finding Algorithm", "abstract": " We consider an Anonymous Multi-Agent Path-Finding (AMAPF) problem where the\nset of agents is confined to a graph, a set of goal vertices is given and each\nof these vertices has to be reached by some agent. The problem is to find an\nassignment of the goals to the agents as well as the collision-free paths, and\nwe are interested in finding the solution with the optimal makespan. A\nwell-established approach to solve this problem is to reduce it to a special\ntype of a graph search problem, i.e. to the problem of finding a maximum flow\non an auxiliary graph induced by the input one. The size of the former graph\nmay be very large and the search on it may become a bottleneck. To this end, we\nsuggest a specific search algorithm that leverages the idea of exploring the\nsearch space not through considering separate search states but rather bulks of\nthem simultaneously. That is, we implicitly compress, store and expand bulks of\nthe search states as single states, which results in high reduction in runtime\nand memory. Empirically, the resultant AMAPF solver demonstrates superior\nperformance compared to the state-of-the-art competitor and is able to solve\nall publicly available MAPF instances from the well-known MovingAI benchmark in\nless than 30 seconds.\n", "date": "2023-12-17", "categories": []} {"id": "2211.14409", "title": "Domain-Independent Dynamic Programming: Generic State Space Search for\n Combinatorial Optimization", "abstract": " For combinatorial optimization problems, model-based approaches such as\nmixed-integer programming (MIP) and constraint programming (CP) aim to decouple\nmodeling and solving a problem: the 'holy grail' of declarative problem\nsolving. We propose domain-independent dynamic programming (DIDP), a new\nmodel-based paradigm based on dynamic programming (DP). While DP is not new, it\nhas typically been implemented as a problem-specific method. We propose Dynamic\nProgramming Description Language (DyPDL), a formalism to define DP models, and\ndevelop Cost-Algebraic A* Solver for DyPDL (CAASDy), a generic solver for DyPDL\nusing state space search. We formalize existing problem-specific DP and state\nspace search methods for combinatorial optimization problems as DP models in\nDyPDL. Using CAASDy and commercial MIP and CP solvers, we experimentally\ncompare the DP models with existing MIP and CP models, showing that, despite\nits nascent nature, CAASDy outperforms MIP and CP on a number of common problem\nclasses.\n", "date": "2022-11-26", "categories": []} {"id": "2307.12547", "title": "Knapsack: Connectedness, Path, and Shortest-Path", "abstract": " We study the knapsack problem with graph theoretic constraints. That is, we\nassume that there exists a graph structure on the set of items of knapsack and\nthe solution also needs to satisfy certain graph theoretic properties on top of\nknapsack constraints. In particular, we need to compute in the connected\nknapsack problem a connected subset of items which has maximum value subject to\nthe size of knapsack constraint. We show that this problem is strongly\nNP-complete even for graphs of maximum degree four and NP-complete even for\nstar graphs. On the other hand, we develop an algorithm running in time\n$O\\left(2^{tw\\log tw}\\cdot\\text{poly}(\\min\\{s^2,d^2\\})\\right)$ where $tw,s,d$\nare respectively treewidth of the graph, size, and target value of the\nknapsack. We further exhibit a $(1-\\epsilon)$ factor approximation algorithm\nrunning in time $O\\left(2^{tw\\log tw}\\cdot\\text{poly}(n,1/\\epsilon)\\right)$ for\nevery $\\epsilon>0$. We show similar results for several other graph theoretic\nproperties, namely path and shortest-path under the problem names path-knapsack\nand shortestpath-knapsack. Our results seems to indicate that\nconnected-knapsack is computationally hardest followed by path-knapsack and\nshortestpath-knapsack.\n", "date": "2023-07-24", "categories": []} {"id": "2308.11477", "title": "An improved column-generation-based matheuristic for learning\n classification trees", "abstract": " Decision trees are highly interpretable models for solving classification\nproblems in machine learning (ML). The standard ML algorithms for training\ndecision trees are fast but generate suboptimal trees in terms of accuracy.\nOther discrete optimization models in the literature address the optimality\nproblem but only work well on relatively small datasets. \\cite{firat2020column}\nproposed a column-generation-based heuristic approach for learning decision\ntrees. This approach improves scalability and can work with large datasets. In\nthis paper, we describe improvements to this column generation approach. First,\nwe modify the subproblem model to significantly reduce the number of\nsubproblems in multiclass classification instances. Next, we show that the\ndata-dependent constraints in the master problem are implied, and use them as\ncutting planes. Furthermore, we describe a separation model to generate data\npoints for which the linear programming relaxation solution violates their\ncorresponding constraints. We conclude by presenting computational results that\nshow that these modifications result in better scalability.\n", "date": "2023-08-22", "categories": []} {"id": "2308.04292", "title": "Engineering LaCAM$^\\ast$: Towards Real-Time, Large-Scale, and\n Near-Optimal Multi-Agent Pathfinding", "abstract": " This paper addresses the challenges of real-time, large-scale, and\nnear-optimal multi-agent pathfinding (MAPF) through enhancements to the\nrecently proposed LaCAM* algorithm. LaCAM* is a scalable search-based algorithm\nthat guarantees the eventual finding of optimal solutions for cumulative\ntransition costs. While it has demonstrated remarkable planning success rates,\nsurpassing various state-of-the-art MAPF methods, its initial solution quality\nis far from optimal, and its convergence speed to the optimum is slow. To\novercome these limitations, this paper introduces several improvement\ntechniques, partly drawing inspiration from other MAPF methods. We provide\nempirical evidence that the fusion of these techniques significantly improves\nthe solution quality of LaCAM*, thus further pushing the boundaries of MAPF\nalgorithms.\n", "date": "2023-08-08", "categories": []} {"id": "2401.10965", "title": "Decentralizing Coordination in Open Vehicle Fleets for Scalable and\n Dynamic Task Allocation", "abstract": " One of the major challenges in the coordination of large, open,\ncollaborative, and commercial vehicle fleets is dynamic task allocation.\nSelf-concerned individually rational vehicle drivers have both local and global\nobjectives, which require coordination using some fair and efficient task\nallocation method. In this paper, we review the literature on scalable and\ndynamic task allocation focusing on deterministic and dynamic two-dimensional\nlinear assignment problems. We focus on multiagent system representation of\nopen vehicle fleets where dynamically appearing vehicles are represented by\nsoftware agents that should be allocated to a set of dynamically appearing\ntasks. We give a comparison and critical analysis of recent research results\nfocusing on centralized, distributed, and decentralized solution approaches.\nMoreover, we propose mathematical models for dynamic versions of the following\nassignment problems well known in combinatorial optimization: the assignment\nproblem, bottleneck assignment problem, fair matching problem, dynamic minimum\ndeviation assignment problem, $\\sum_{k}$-assignment problem, the semiassignment\nproblem, the assignment problem with side constraints, and the assignment\nproblem while recognizing agent qualification; all while considering the main\naspect of open vehicle fleets: random arrival of tasks and vehicles (agents)\nthat may become available after assisting previous tasks or by participating in\nthe fleet at times based on individual interest.\n", "date": "2024-01-19", "categories": []} {"id": "2401.10274", "title": "Knowledge-Assisted Dual-Stage Evolutionary Optimization of Large-Scale\n Crude Oil Scheduling", "abstract": " With the scaling up of crude oil scheduling in modern refineries, large-scale\ncrude oil scheduling problems (LSCOSPs) emerge with thousands of binary\nvariables and non-linear constraints, which are challenging to be optimized by\ntraditional optimization methods. To solve LSCOSPs, we take the practical crude\noil scheduling from a marine-access refinery as an example and start with\nmodeling LSCOSPs from crude unloading, transportation, crude distillation unit\nprocessing, and inventory management of intermediate products. On the basis of\nthe proposed model, a dual-stage evolutionary algorithm driven by heuristic\nrules (denoted by DSEA/HR) is developed, where the dual-stage search mechanism\nconsists of global search and local refinement. In the global search stage, we\ndevise several heuristic rules based on the empirical operating knowledge to\ngenerate a well-performing initial population and accelerate convergence in the\nmixed variables space. In the local refinement stage, a repair strategy is\nproposed to move the infeasible solutions towards feasible regions by further\noptimizing the local continuous variables. During the whole evolutionary\nprocess, the proposed dual-stage framework plays a crucial role in balancing\nexploration and exploitation. Experimental results have shown that DSEA/HR\noutperforms the state-of-the-art and widely-used mathematical programming\nmethods and metaheuristic algorithms on LSCOSP instances within a reasonable\ntime.\n", "date": "2024-01-09", "categories": []} {"id": "2401.10420", "title": "Generalized Nested Rollout Policy Adaptation with Limited Repetitions", "abstract": " Generalized Nested Rollout Policy Adaptation (GNRPA) is a Monte Carlo search\nalgorithm for optimizing a sequence of choices. We propose to improve on GNRPA\nby avoiding too deterministic policies that find again and again the same\nsequence of choices. We do so by limiting the number of repetitions of the best\nsequence found at a given level. Experiments show that it improves the\nalgorithm for three different combinatorial problems: Inverse RNA Folding, the\nTraveling Salesman Problem with Time Windows and the Weak Schur problem.\n", "date": "2024-01-18", "categories": []} {"id": "2401.10431", "title": "Learning a Prior for Monte Carlo Search by Replaying Solutions to\n Combinatorial Problems", "abstract": " Monte Carlo Search gives excellent results in multiple difficult\ncombinatorial problems. Using a prior to perform non uniform playouts during\nthe search improves a lot the results compared to uniform playouts. Handmade\nheuristics tailored to the combinatorial problem are often used as priors. We\npropose a method to automatically compute a prior. It uses statistics on solved\nproblems. It is a simple and general method that incurs no computational cost\nat playout time and that brings large performance gains. The method is applied\nto three difficult combinatorial problems: Latin Square Completion, Kakuro, and\nInverse RNA Folding.\n", "date": "2024-01-19", "categories": []} {"id": "2401.10589", "title": "Rethinking the Soft Conflict Pseudo Boolean Constraint on MaxSAT Local\n Search Solvers", "abstract": " MaxSAT is an optimization version of the famous NP-complete Satisfiability\nproblem (SAT). Algorithms for MaxSAT mainly include complete solvers and local\nsearch incomplete solvers. In many complete solvers, once a better solution is\nfound, a Soft conflict Pseudo Boolean (SPB) constraint will be generated to\nenforce the algorithm to find better solutions. In many local search\nalgorithms, clause weighting is a key technique for effectively guiding the\nsearch directions. In this paper, we propose to transfer the SPB constraint\ninto the clause weighting system of the local search method, leading the\nalgorithm to better solutions. We further propose an adaptive clause weighting\nstrategy that breaks the tradition of using constant values to adjust clause\nweights. Based on the above methods, we propose a new local search algorithm\ncalled SPB-MaxSAT that provides new perspectives for clause weighting on MaxSAT\nlocal search solvers. Extensive experiments demonstrate the excellent\nperformance of the proposed methods.\n", "date": "2024-01-19", "categories": []} {"id": "2211.13118", "title": "Decision Diagram-Based Branch-and-Bound with Caching for Dominance and\n Suboptimality Detection", "abstract": " The branch-and-bound algorithm based on decision diagrams introduced by\nBergman et al. in 2016 is a framework for solving discrete optimization\nproblems with a dynamic programming formulation. It works by compiling a series\nof bounded-width decision diagrams that can provide lower and upper bounds for\nany given subproblem. Eventually, every part of the search space will be either\nexplored or pruned by the algorithm, thus proving optimality. This paper\npresents new ingredients to speed up the search by exploiting the structure of\ndynamic programming models. The key idea is to prevent the repeated expansion\nof nodes corresponding to the same dynamic programming states by querying\nexpansion thresholds cached throughout the search. These thresholds are based\non dominance relations between partial solutions previously found and on the\npruning inequalities of the filtering techniques introduced by Gillard et al.\nin 2021. Computational experiments show that the pruning brought by this\ncaching mechanism allows significantly reducing the number of nodes expanded by\nthe algorithm. This results in more benchmark instances of difficult\noptimization problems being solved in less time while using narrower decision\ndiagrams.\n", "date": "2022-11-22", "categories": []} {"id": "2312.12469", "title": "Distilling Autoregressive Models to Obtain High-Performance\n Non-Autoregressive Solvers for Vehicle Routing Problems with Faster Inference\n Speed", "abstract": " Neural construction models have shown promising performance for Vehicle\nRouting Problems (VRPs) by adopting either the Autoregressive (AR) or\nNon-Autoregressive (NAR) learning approach. While AR models produce\nhigh-quality solutions, they generally have a high inference latency due to\ntheir sequential generation nature. Conversely, NAR models generate solutions\nin parallel with a low inference latency but generally exhibit inferior\nperformance. In this paper, we propose a generic Guided Non-Autoregressive\nKnowledge Distillation (GNARKD) method to obtain high-performance NAR models\nhaving a low inference latency. GNARKD removes the constraint of sequential\ngeneration in AR models while preserving the learned pivotal components in the\nnetwork architecture to obtain the corresponding NAR models through knowledge\ndistillation. We evaluate GNARKD by applying it to three widely adopted AR\nmodels to obtain NAR VRP solvers for both synthesized and real-world instances.\nThe experimental results demonstrate that GNARKD significantly reduces the\ninference time (4-5 times faster) with acceptable performance drop (2-3\\%). To\nthe best of our knowledge, this study is first-of-its-kind to obtain NAR VRP\nsolvers from AR ones through knowledge distillation.\n", "date": "2023-12-19", "categories": []} {"id": "2311.05937", "title": "Genetic Algorithm enhanced by Deep Reinforcement Learning in parent\n selection mechanism and mutation : Minimizing makespan in permutation flow\n shop scheduling problems", "abstract": " This paper introduces a reinforcement learning (RL) approach to address the\nchallenges associated with configuring and optimizing genetic algorithms (GAs)\nfor solving difficult combinatorial or non-linear problems. The proposed RL+GA\nmethod was specifically tested on the flow shop scheduling problem (FSP). The\nhybrid algorithm incorporates neural networks (NN) and uses the off-policy\nmethod Q-learning or the on-policy method Sarsa(0) to control two key genetic\nalgorithm (GA) operators: parent selection mechanism and mutation. At each\ngeneration, the RL agent's action is determining the selection method, the\nprobability of the parent selection and the probability of the offspring\nmutation. This allows the RL agent to dynamically adjust the selection and\nmutation based on its learned policy. The results of the study highlight the\neffectiveness of the RL+GA approach in improving the performance of the\nprimitive GA. They also demonstrate its ability to learn and adapt from\npopulation diversity and solution improvements over time. This adaptability\nleads to improved scheduling solutions compared to static parameter\nconfigurations while maintaining population diversity throughout the\nevolutionary process.\n", "date": "2023-11-10", "categories": []} {"id": "2401.05960", "title": "Machine Learning Insides OptVerse AI Solver: Design Principles and\n Applications", "abstract": " In an era of digital ubiquity, efficient resource management and\ndecision-making are paramount across numerous industries. To this end, we\npresent a comprehensive study on the integration of machine learning (ML)\ntechniques into Huawei Cloud's OptVerse AI Solver, which aims to mitigate the\nscarcity of real-world mathematical programming instances, and to surpass the\ncapabilities of traditional optimization techniques. We showcase our methods\nfor generating complex SAT and MILP instances utilizing generative models that\nmirror multifaceted structures of real-world problem. Furthermore, we introduce\na training framework leveraging augmentation policies to maintain solvers'\nutility in dynamic environments. Besides the data generation and augmentation,\nour proposed approaches also include novel ML-driven policies for personalized\nsolver strategies, with an emphasis on applications like graph convolutional\nnetworks for initial basis selection and reinforcement learning for advanced\npresolving and cut selection. Additionally, we detail the incorporation of\nstate-of-the-art parameter tuning algorithms which markedly elevate solver\nperformance. Compared with traditional solvers such as Cplex and SCIP, our\nML-augmented OptVerse AI Solver demonstrates superior speed and precision\nacross both established benchmarks and real-world scenarios, reinforcing the\npractical imperative and effectiveness of machine learning techniques in\nmathematical programming solvers.\n", "date": "2024-01-11", "categories": []} {"id": "2302.01578", "title": "Searching Large Neighborhoods for Integer Linear Programs with\n Contrastive Learning", "abstract": " Integer Linear Programs (ILPs) are powerful tools for modeling and solving a\nlarge number of combinatorial optimization problems. Recently, it has been\nshown that Large Neighborhood Search (LNS), as a heuristic algorithm, can find\nhigh quality solutions to ILPs faster than Branch and Bound. However, how to\nfind the right heuristics to maximize the performance of LNS remains an open\nproblem. In this paper, we propose a novel approach, CL-LNS, that delivers\nstate-of-the-art anytime performance on several ILP benchmarks measured by\nmetrics including the primal gap, the primal integral, survival rates and the\nbest performing rate. Specifically, CL-LNS collects positive and negative\nsolution samples from an expert heuristic that is slow to compute and learns a\nnew one with a contrastive loss. We use graph attention networks and a richer\nset of features to further improve its performance.\n", "date": "2023-02-03", "categories": []} {"id": "2305.02209", "title": "Large-scale Online Ridesharing: The Effect of Assignment Optimality on\n System Performance", "abstract": " Mobility-on-demand (MoD) systems consist of a fleet of shared vehicles that\ncan be hailed for one-way point-to-point trips. The total distance driven by\nthe vehicles and the fleet size can be reduced by employing ridesharing, i.e.,\nby assigning multiple passengers to one vehicle. However, finding the optimal\npassenger-vehicle assignment in an MoD system is a hard combinatorial problem.\nIn this work, we demonstrate how the VGA method, a recently proposed systematic\nmethod for ridesharing, can be used to compute the optimal passenger-vehicle\nassignments and corresponding vehicle routes in a massive-scale MoD system. In\ncontrast to existing works, we solve all passenger-vehicle assignment problems\nto optimality, regularly dealing with instances containing thousands of\nvehicles and passengers. Moreover, to examine the impact of using optimal\nridesharing assignments, we compare the performance of an MoD system that uses\noptimal assignments against an MoD system that uses assignments computed using\ninsertion heuristic and against an MoD system that uses no ridesharing. We\nfound that the system that uses optimal ridesharing assignments subject to the\nmaximum travel delay of 4 minutes reduces the vehicle distance driven by 57 %\ncompared to an MoD system without ridesharing. Furthermore, we found that the\noptimal assignments result in a 20 % reduction in vehicle distance driven and 5\n% lower average passenger travel delay compared to a system that uses insertion\nheuristic.\n", "date": "2023-04-07", "categories": []} {"id": "2309.12678", "title": "QAL-BP: An Augmented Lagrangian Quantum Approach for Bin Packing", "abstract": " The bin packing is a well-known NP-Hard problem in the domain of artificial\nintelligence, posing significant challenges in finding efficient solutions.\nConversely, recent advancements in quantum technologies have shown promising\npotential for achieving substantial computational speedup, particularly in\ncertain problem classes, such as combinatorial optimization. In this study, we\nintroduce QAL-BP, a novel Quadratic Unconstrained Binary Optimization (QUBO)\nformulation designed specifically for bin packing and suitable for quantum\ncomputation. QAL-BP utilizes the Augmented Lagrangian method to incorporate the\nbin packing constraints into the objective function while also facilitating an\nanalytical estimation of heuristic, but empirically robust, penalty\nmultipliers. This approach leads to a more versatile and generalizable model\nthat eliminates the need for empirically calculating instance-dependent\nLagrangian coefficients, a requirement commonly encountered in alternative QUBO\nformulations for similar problems. To assess the effectiveness of our proposed\napproach, we conduct experiments on a set of bin packing instances using a real\nQuantum Annealing device. Additionally, we compare the results with those\nobtained from two different classical solvers, namely simulated annealing and\nGurobi. The experimental findings not only confirm the correctness of the\nproposed formulation but also demonstrate the potential of quantum computation\nin effectively solving the bin packing problem, particularly as more reliable\nquantum technology becomes available.\n", "date": "2023-09-22", "categories": []} {"id": "2310.07985", "title": "Neural Combinatorial Optimization with Heavy Decoder: Toward Large Scale\n Generalization", "abstract": " Neural combinatorial optimization (NCO) is a promising learning-based\napproach for solving challenging combinatorial optimization problems without\nspecialized algorithm design by experts. However, most constructive NCO methods\ncannot solve problems with large-scale instance sizes, which significantly\ndiminishes their usefulness for real-world applications. In this work, we\npropose a novel Light Encoder and Heavy Decoder (LEHD) model with a strong\ngeneralization ability to address this critical issue. The LEHD model can learn\nto dynamically capture the relationships between all available nodes of varying\nsizes, which is beneficial for model generalization to problems of various\nscales. Moreover, we develop a data-efficient training scheme and a flexible\nsolution construction mechanism for the proposed LEHD model. By training on\nsmall-scale problem instances, the LEHD model can generate nearly optimal\nsolutions for the Travelling Salesman Problem (TSP) and the Capacitated Vehicle\nRouting Problem (CVRP) with up to 1000 nodes, and also generalizes well to\nsolve real-world TSPLib and CVRPLib problems. These results confirm our\nproposed LEHD model can significantly improve the state-of-the-art performance\nfor constructive NCO. The code is available at\nhttps://github.com/CIAM-Group/NCO_code/tree/main/single_objective/LEHD.\n", "date": "2023-10-12", "categories": []} {"id": "2401.06834", "title": "Optimization of Discrete Parameters Using the Adaptive Gradient Method\n and Directed Evolution", "abstract": " The problem is considered of optimizing discrete parameters in the presence\nof constraints. We use the stochastic sigmoid with temperature and put forward\nthe new adaptive gradient method CONGA. The search for an optimal solution is\ncarried out by a population of individuals. Each of them varies according to\ngradients of the 'environment' and is characterized by two temperature\nparameters with different annealing schedules. Unadapted individuals die, and\noptimal ones interbreed, the result is directed evolutionary dynamics. The\nproposed method is illustrated using the well-known combinatorial problem for\noptimal packing of a backpack (0-1 KP).\n", "date": "2024-01-12", "categories": []} {"id": "2401.06952", "title": "Reinforcement Learning for Scalable Train Timetable Rescheduling with\n Graph Representation", "abstract": " Train timetable rescheduling (TTR) aims to promptly restore the original\noperation of trains after unexpected disturbances or disruptions. Currently,\nthis work is still done manually by train dispatchers, which is challenging to\nmaintain performance under various problem instances. To mitigate this issue,\nthis study proposes a reinforcement learning-based approach to TTR, which makes\nthe following contributions compared to existing work. First, we design a\nsimple directed graph to represent the TTR problem, enabling the automatic\nextraction of informative states through graph neural networks. Second, we\nreformulate the construction process of TTR's solution, not only decoupling the\ndecision model from the problem size but also ensuring the generated scheme's\nfeasibility. Third, we design a learning curriculum for our model to handle the\nscenarios with different levels of delay. Finally, a simple local search method\nis proposed to assist the learned decision model, which can significantly\nimprove solution quality with little additional computation cost, further\nenhancing the practical value of our method. Extensive experimental results\ndemonstrate the effectiveness of our method. The learned decision model can\nachieve better performance for various problems with varying degrees of train\ndelay and different scales when compared to handcrafted rules and\nstate-of-the-art solvers.\n", "date": "2024-01-13", "categories": []} {"id": "2401.06979", "title": "Distance-aware Attention Reshaping: Enhance Generalization of Neural\n Solver for Large-scale Vehicle Routing Problems", "abstract": " Neural solvers based on attention mechanism have demonstrated remarkable\neffectiveness in solving vehicle routing problems. However, in the\ngeneralization process from small scale to large scale, we find a phenomenon of\nthe dispersion of attention scores in existing neural solvers, which leads to\npoor performance. To address this issue, this paper proposes a distance-aware\nattention reshaping method, assisting neural solvers in solving large-scale\nvehicle routing problems. Specifically, without the need for additional\ntraining, we utilize the Euclidean distance information between current nodes\nto adjust attention scores. This enables a neural solver trained on small-scale\ninstances to make rational choices when solving a large-scale problem.\nExperimental results show that the proposed method significantly outperforms\nexisting state-of-the-art neural solvers on the large-scale CVRPLib dataset.\n", "date": "2024-01-13", "categories": []} {"id": "2401.05610", "title": "Graph Q-Learning for Combinatorial Optimization", "abstract": " Graph-structured data is ubiquitous throughout natural and social sciences,\nand Graph Neural Networks (GNNs) have recently been shown to be effective at\nsolving prediction and inference problems on graph data. In this paper, we\npropose and demonstrate that GNNs can be applied to solve Combinatorial\nOptimization (CO) problems. CO concerns optimizing a function over a discrete\nsolution space that is often intractably large. To learn to solve CO problems,\nwe formulate the optimization process as a sequential decision making problem,\nwhere the return is related to how close the candidate solution is to\noptimality. We use a GNN to learn a policy to iteratively build increasingly\npromising candidate solutions. We present preliminary evidence that GNNs\ntrained through Q-Learning can solve CO problems with performance approaching\nstate-of-the-art heuristic-based solvers, using only a fraction of the\nparameters and training time.\n", "date": "2024-01-11", "categories": []} {"id": "2401.05969", "title": "Spatial-Aware Deep Reinforcement Learning for the Traveling Officer\n Problem", "abstract": " The traveling officer problem (TOP) is a challenging stochastic optimization\ntask. In this problem, a parking officer is guided through a city equipped with\nparking sensors to fine as many parking offenders as possible. A major\nchallenge in TOP is the dynamic nature of parking offenses, which randomly\nappear and disappear after some time, regardless of whether they have been\nfined. Thus, solutions need to dynamically adjust to currently fineable parking\noffenses while also planning ahead to increase the likelihood that the officer\narrives during the offense taking place. Though various solutions exist, these\nmethods often struggle to take the implications of actions on the ability to\nfine future parking violations into account. This paper proposes SATOP, a novel\nspatial-aware deep reinforcement learning approach for TOP. Our novel state\nencoder creates a representation of each action, leveraging the spatial\nrelationships between parking spots, the agent, and the action. Furthermore, we\npropose a novel message-passing module for learning future inter-action\ncorrelations in the given environment. Thus, the agent can estimate the\npotential to fine further parking violations after executing an action. We\nevaluate our method using an environment based on real-world data from\nMelbourne. Our results show that SATOP consistently outperforms\nstate-of-the-art TOP agents and is able to fine up to 22% more parking\noffenses.\n", "date": "2024-01-11", "categories": []} {"id": "2307.15557", "title": "Dynamic algorithms for k-center on graphs", "abstract": " In this paper we give the first efficient algorithms for the $k$-center\nproblem on dynamic graphs undergoing edge updates. In this problem, the goal is\nto partition the input into $k$ sets by choosing $k$ centers such that the\nmaximum distance from any data point to its closest center is minimized. It is\nknown that it is NP-hard to get a better than $2$ approximation for this\nproblem.\n While in many applications the input may naturally be modeled as a graph, all\nprior works on $k$-center problem in dynamic settings are on point sets in\narbitrary metric spaces. In this paper, we give a deterministic decremental\n$(2+\\epsilon)$-approximation algorithm and a randomized incremental\n$(4+\\epsilon)$-approximation algorithm, both with amortized update time\n$kn^{o(1)}$ for weighted graphs. Moreover, we show a reduction that leads to a\nfully dynamic $(2+\\epsilon)$-approximation algorithm for the $k$-center\nproblem, with worst-case update time that is within a factor $k$ of the\nstate-of-the-art fully dynamic $(1+\\epsilon)$-approximation single-source\nshortest paths algorithm in graphs. Matching this bound is a natural goalpost\nbecause the approximate distances of each vertex to its center can be used to\nmaintain a $(2+\\epsilon)$-approximation of the graph diameter and the fastest\nknown algorithms for such a diameter approximation also rely on maintaining\napproximate single-source distances.\n", "date": "2023-07-28", "categories": []} {"id": "2309.04339", "title": "Online Submodular Maximization via Online Convex Optimization", "abstract": " We study monotone submodular maximization under general matroid constraints\nin the online setting. We prove that online optimization of a large class of\nsubmodular functions, namely, weighted threshold potential functions, reduces\nto online convex optimization (OCO). This is precisely because functions in\nthis class admit a concave relaxation; as a result, OCO policies, coupled with\nan appropriate rounding scheme, can be used to achieve sublinear regret in the\ncombinatorial setting. We show that our reduction extends to many different\nversions of the online learning problem, including the dynamic regret, bandit,\nand optimistic-learning settings.\n", "date": "2023-09-08", "categories": []} {"id": "2311.00983", "title": "Optimizing Inventory Routing: A Decision-Focused Learning Approach using\n Neural Networks", "abstract": " Inventory Routing Problem (IRP) is a crucial challenge in supply chain\nmanagement as it involves optimizing efficient route selection while\nconsidering the uncertainty of inventory demand planning. To solve IRPs,\nusually a two-stage approach is employed, where demand is predicted using\nmachine learning techniques first, and then an optimization algorithm is used\nto minimize routing costs. Our experiment shows machine learning models fall\nshort of achieving perfect accuracy because inventory levels are influenced by\nthe dynamic business environment, which, in turn, affects the optimization\nproblem in the next stage, resulting in sub-optimal decisions. In this paper,\nwe formulate and propose a decision-focused learning-based approach to solving\nreal-world IRPs. This approach directly integrates inventory prediction and\nrouting optimization within an end-to-end system potentially ensuring a robust\nsupply chain strategy.\n", "date": "2023-11-02", "categories": []} {"id": "2312.09086", "title": "COMBHelper: A Neural Approach to Reduce Search Space for Graph\n Combinatorial Problems", "abstract": " Combinatorial Optimization (CO) problems over graphs appear routinely in many\napplications such as in optimizing traffic, viral marketing in social networks,\nand matching for job allocation. Due to their combinatorial nature, these\nproblems are often NP-hard. Existing approximation algorithms and heuristics\nrely on the search space to find the solutions and become time-consuming when\nthis space is large. In this paper, we design a neural method called COMBHelper\nto reduce this space and thus improve the efficiency of the traditional CO\nalgorithms based on node selection. Specifically, it employs a Graph Neural\nNetwork (GNN) to identify promising nodes for the solution set. This pruned\nsearch space is then fed to the traditional CO algorithms. COMBHelper also uses\na Knowledge Distillation (KD) module and a problem-specific boosting module to\nbring further efficiency and efficacy. Our extensive experiments show that the\ntraditional CO algorithms with COMBHelper are at least 2 times faster than\ntheir original versions.\n", "date": "2023-12-14", "categories": []} {"id": "2312.16767", "title": "Adaptive Anytime Multi-Agent Path Finding Using Bandit-Based Large\n Neighborhood Search", "abstract": " Anytime multi-agent path finding (MAPF) is a promising approach to scalable\npath optimization in large-scale multi-agent systems. State-of-the-art anytime\nMAPF is based on Large Neighborhood Search (LNS), where a fast initial solution\nis iteratively optimized by destroying and repairing a fixed number of parts,\ni.e., the neighborhood, of the solution, using randomized destroy heuristics\nand prioritized planning. Despite their recent success in various MAPF\ninstances, current LNS-based approaches lack exploration and flexibility due to\ngreedy optimization with a fixed neighborhood size which can lead to low\nquality solutions in general. So far, these limitations have been addressed\nwith extensive prior effort in tuning or offline machine learning beyond actual\nplanning. In this paper, we focus on online learning in LNS and propose\nBandit-based Adaptive LArge Neighborhood search Combined with Exploration\n(BALANCE). BALANCE uses a bi-level multi-armed bandit scheme to adapt the\nselection of destroy heuristics and neighborhood sizes on the fly during\nsearch. We evaluate BALANCE on multiple maps from the MAPF benchmark set and\nempirically demonstrate cost improvements of at least 50% compared to\nstate-of-the-art anytime MAPF in large-scale scenarios. We find that Thompson\nSampling performs particularly well compared to alternative multi-armed bandit\nalgorithms.\n", "date": "2023-12-28", "categories": []} {"id": "2312.16154", "title": "Clustered Orienteering Problem with Subgroups", "abstract": " This paper introduces an extension to the Orienteering Problem (OP), called\nClustered Orienteering Problem with Subgroups (COPS). In this variant, nodes\nare arranged into subgroups, and the subgroups are organized into clusters. A\nreward is associated with each subgroup and is gained only if all of its nodes\nare visited; however, at most one subgroup can be visited per cluster. The\nobjective is to maximize the total collected reward while attaining a travel\nbudget. We show that our new formulation has the ability to model and solve two\nprevious well-known variants, the Clustered Orienteering Problem (COP) and the\nSet Orienteering Problem (SOP), in addition to other scenarios introduced here.\nAn Integer Linear Programming (ILP) formulation and a Tabu Search-based\nheuristic are proposed to solve the problem. Experimental results indicate that\nthe ILP method can yield optimal solutions at the cost of time, whereas the\nmetaheuristic produces comparable solutions within a more reasonable\ncomputational cost.\n", "date": "2023-12-26", "categories": []} {"id": "2312.16423", "title": "General Method for Solving Four Types of SAT Problems", "abstract": " Existing methods provide varying algorithms for different types of Boolean\nsatisfiability problems (SAT), lacking a general solution framework.\nAccordingly, this study proposes a unified framework DCSAT based on integer\nprogramming and reinforcement learning (RL) algorithm to solve different types\nof SAT problems such as MaxSAT, Weighted MaxSAT, PMS, WPMS. Specifically, we\nfirst construct a consolidated integer programming representation for four\ntypes of SAT problems by adjusting objective function coefficients. Secondly,\nwe construct an appropriate reinforcement learning models based on the 0-1\ninteger programming for SAT problems. Based on the binary tree search\nstructure, we apply the Monte Carlo tree search (MCTS) method on SAT problems.\nFinally, we prove that this method can find all optimal Boolean assignments\nbased on Wiener-khinchin law of large Numbers. We experimentally verify that\nthis paradigm can prune the unnecessary search space to find the optimal\nBoolean assignments for the problem. Furthermore, the proposed method can\nprovide diverse labels for supervised learning methods for SAT problems.\n", "date": "2023-12-27", "categories": []} {"id": "2306.00959", "title": "Dynamic Algorithms for Matroid Submodular Maximization", "abstract": " Submodular maximization under matroid and cardinality constraints are\nclassical problems with a wide range of applications in machine learning,\nauction theory, and combinatorial optimization. In this paper, we consider\nthese problems in the dynamic setting, where (1) we have oracle access to a\nmonotone submodular function $f: 2^{V} \\rightarrow \\mathbb{R}^+$ and (2) we are\ngiven a sequence $\\mathcal{S}$ of insertions and deletions of elements of an\nunderlying ground set $V$.\n We develop the first fully dynamic $(4+\\epsilon)$-approximation algorithm for\nthe submodular maximization problem under the matroid constraint using an\nexpected worst-case $O(k\\log(k)\\log^3{(k/\\epsilon)})$ query complexity where $0\n< \\epsilon \\le 1$. This resolves an open problem of Chen and Peng (STOC'22) and\nLattanzi et al. (NeurIPS'20).\n As a byproduct, for the submodular maximization under the cardinality\nconstraint $k$, we propose a parameterized (by the cardinality constraint $k$)\ndynamic algorithm that maintains a $(2+\\epsilon)$-approximate solution of the\nsequence $\\mathcal{S}$ at any time $t$ using an expected worst-case query\ncomplexity $O(k\\epsilon^{-1}\\log^2(k))$. This is the first dynamic algorithm\nfor the problem that has a query complexity independent of the size of ground\nset $V$.\n", "date": "2023-06-01", "categories": []} {"id": "2312.15412", "title": "CARSS: Cooperative Attention-guided Reinforcement Subpath Synthesis for\n Solving Traveling Salesman Problem", "abstract": " This paper introduces CARSS (Cooperative Attention-guided Reinforcement\nSubpath Synthesis), a novel approach to address the Traveling Salesman Problem\n(TSP) by leveraging cooperative Multi-Agent Reinforcement Learning (MARL).\nCARSS decomposes the TSP solving process into two distinct yet synergistic\nsteps: \"subpath generation\" and \"subpath merging.\" In the former, a cooperative\nMARL framework is employed to iteratively generate subpaths using multiple\nagents. In the latter, these subpaths are progressively merged to form a\ncomplete cycle. The algorithm's primary objective is to enhance efficiency in\nterms of training memory consumption, testing time, and scalability, through\nthe adoption of a multi-agent divide and conquer paradigm. Notably, attention\nmechanisms play a pivotal role in feature embedding and parameterization\nstrategies within CARSS. The training of the model is facilitated by the\nindependent REINFORCE algorithm. Empirical experiments reveal CARSS's\nsuperiority compared to single-agent alternatives: it demonstrates reduced GPU\nmemory utilization, accommodates training graphs nearly 2.5 times larger, and\nexhibits the potential for scaling to even more extensive problem sizes.\nFurthermore, CARSS substantially reduces testing time and optimization gaps by\napproximately 50% for TSP instances of up to 1000 vertices, when compared to\nstandard decoding methods.\n", "date": "2023-12-24", "categories": []} {"id": "2312.15658", "title": "Swap-based Deep Reinforcement Learning for Facility Location Problems in\n Networks", "abstract": " Facility location problems on graphs are ubiquitous in real world and hold\nsignificant importance, yet their resolution is often impeded by NP-hardness.\nRecently, machine learning methods have been proposed to tackle such classical\nproblems, but they are limited to the myopic constructive pattern and only\nconsider the problems in Euclidean space. To overcome these limitations, we\npropose a general swap-based framework that addresses the p-median problem and\nthe facility relocation problem on graphs and a novel reinforcement learning\nmodel demonstrating a keen awareness of complex graph structures. Striking a\nharmonious balance between solution quality and running time, our method\nsurpasses handcrafted heuristics on intricate graph datasets. Additionally, we\nintroduce a graph generation process to simulate real-world urban road networks\nwith demand, facilitating the construction of large datasets for the classic\nproblem. For the initialization of the locations of facilities, we introduce a\nphysics-inspired strategy for the p-median problem, reaching more stable\nsolutions than the random strategy. The proposed pipeline coupling the classic\nswap-based method with deep reinforcement learning marks a significant step\nforward in addressing the practical challenges associated with facility\nlocation on graphs.\n", "date": "2023-12-25", "categories": []} {"id": "2312.15864", "title": "BalMCTS: Balancing Objective Function and Search Nodes in MCTS for\n Constraint Optimization Problems", "abstract": " Constraint Optimization Problems (COP) pose intricate challenges in\ncombinatorial problems usually addressed through Branch and Bound (B\\&B)\nmethods, which involve maintaining priority queues and iteratively selecting\nbranches to search for solutions. However, conventional approaches take a\nconsiderable amount of time to find optimal solutions, and it is also crucial\nto quickly identify a near-optimal feasible solution in a shorter time. In this\npaper, we aim to investigate the effectiveness of employing a depth-first\nsearch algorithm for solving COP, specifically focusing on identifying optimal\nor near-optimal solutions within top $n$ solutions. Hence, we propose a novel\nheuristic neural network algorithm based on MCTS, which, by simultaneously\nconducting search and training, enables the neural network to effectively serve\nas a heuristic during Backtracking. Furthermore, our approach incorporates\nencoding COP problems and utilizing graph neural networks to aggregate\ninformation about variables and constraints, offering more appropriate\nvariables for assignments. Experimental results on stochastic COP instances\ndemonstrate that our method identifies feasible solutions with a gap of less\nthan 17.63% within the initial 5 feasible solutions. Moreover, when applied to\nattendant Constraint Satisfaction Problem (CSP) instances, our method exhibits\na remarkable reduction of less than 5% in searching nodes compared to\nstate-of-the-art approaches.\n", "date": "2023-12-26", "categories": []} {"id": "2312.15908", "title": "Decentralized Monte Carlo Tree Search for Partially Observable\n Multi-agent Pathfinding", "abstract": " The Multi-Agent Pathfinding (MAPF) problem involves finding a set of\nconflict-free paths for a group of agents confined to a graph. In typical MAPF\nscenarios, the graph and the agents' starting and ending vertices are known\nbeforehand, allowing the use of centralized planning algorithms. However, in\nthis study, we focus on the decentralized MAPF setting, where the agents may\nobserve the other agents only locally and are restricted in communications with\neach other. Specifically, we investigate the lifelong variant of MAPF, where\nnew goals are continually assigned to the agents upon completion of previous\nones. Drawing inspiration from the successful AlphaZero approach, we propose a\ndecentralized multi-agent Monte Carlo Tree Search (MCTS) method for MAPF tasks.\nOur approach utilizes the agent's observations to recreate the intrinsic Markov\ndecision process, which is then used for planning with a tailored for\nmulti-agent tasks version of neural MCTS. The experimental results show that\nour approach outperforms state-of-the-art learnable MAPF solvers. The source\ncode is available at https://github.com/AIRI-Institute/mats-lp.\n", "date": "2023-12-26", "categories": []} {"id": "2312.16106", "title": "Clique Analysis and Bypassing in Continuous-Time Conflict-Based Search", "abstract": " While the study of unit-cost Multi-Agent Pathfinding (MAPF) problems has been\npopular, many real-world problems require continuous time and costs due to\nvarious movement models. In this context, this paper studies symmetry-breaking\nenhancements for Continuous-Time Conflict-Based Search (CCBS), a solver for\ncontinuous-time MAPF. Resolving conflict symmetries in MAPF can require an\nexponential amount of work. We adapt known enhancements from unit-cost domains\nfor CCBS: bypassing, which resolves cost symmetries and biclique constraints\nwhich resolve spatial conflict symmetries. We formulate a novel combination of\nbiclique constraints with disjoint splitting for spatial conflict symmetries.\nFinally, we show empirically that these enhancements yield a statistically\nsignificant performance improvement versus previous state of the art, solving\nproblems for up to 10% or 20% more agents in the same amount of time on dense\ngraphs.\n", "date": "2023-12-26", "categories": []} {"id": "2312.14299", "title": "Fairness in Submodular Maximization over a Matroid Constraint", "abstract": " Submodular maximization over a matroid constraint is a fundamental problem\nwith various applications in machine learning. Some of these applications\ninvolve decision-making over datapoints with sensitive attributes such as\ngender or race. In such settings, it is crucial to guarantee that the selected\nsolution is fairly distributed with respect to this attribute. Recently,\nfairness has been investigated in submodular maximization under a cardinality\nconstraint in both the streaming and offline settings, however the more general\nproblem with matroid constraint has only been considered in the streaming\nsetting and only for monotone objectives. This work fills this gap. We propose\nvarious algorithms and impossibility results offering different trade-offs\nbetween quality, fairness, and generality.\n", "date": "2023-12-21", "categories": []} {"id": "2312.14836", "title": "Learning Lagrangian Multipliers for the Travelling Salesman Problem", "abstract": " Lagrangian relaxation is a versatile mathematical technique employed to relax\nconstraints in an optimization problem, enabling the generation of dual bounds\nto prove the optimality of feasible solutions and the design of efficient\npropagators in constraint programming (such as the weighted circuit\nconstraint). However, the conventional process of deriving Lagrangian\nmultipliers (e.g., using subgradient methods) is often computationally\nintensive, limiting its practicality for large-scale or time-sensitive\nproblems. To address this challenge, we propose an innovative unsupervised\nlearning approach that harnesses the capabilities of graph neural networks to\nexploit the problem structure, aiming to generate accurate Lagrangian\nmultipliers efficiently. We apply this technique to the well-known Held-Karp\nLagrangian relaxation for the travelling salesman problem. The core idea is to\npredict accurate Lagrangian multipliers and to employ them as a warm start for\ngenerating Held-Karp relaxation bounds. These bounds are subsequently utilized\nto enhance the filtering process carried out by branch-and-bound algorithms. In\ncontrast to much of the existing literature, which primarily focuses on finding\nfeasible solutions, our approach operates on the dual side, demonstrating that\nlearning can also accelerate the proof of optimality. We conduct experiments\nacross various distributions of the metric travelling salesman problem,\nconsidering instances with up to 200 cities. The results illustrate that our\napproach can improve the filtering level of the weighted circuit global\nconstraint, reduce the optimality gap by a factor two for unsolved instances up\nto a timeout, and reduce the execution time for solved instances by 10%.\n", "date": "2023-12-22", "categories": []} {"id": "2312.13682", "title": "A Constraint Programming Model for Scheduling the Unloading of Trains in\n Ports: Extended", "abstract": " In this paper, we propose a model to schedule the next 24 hours of operations\nin a bulk cargo port to unload bulk cargo trains onto stockpiles. It is a\nproblem that includes multiple parts such as splitting long trains into shorter\nones and the routing of bulk material through a configurable network of\nconveyors to the stockpiles. Managing such trains (up to three kilometers long)\nalso requires specialized equipment. The real world nature of the problem\nspecification implies the necessity to manage heterogeneous data. Indeed, when\nnew equipment is added (e.g. dumpers) or a new type of wagon comes in use,\nolder or different equipment will still be in use as well. All these details\nneed to be accounted for. In fact, avoiding a full deadlock of the facility\nafter a new but ineffective schedule is produced. In this paper, we provide a\ndetailed presentation of this real world problem and its associated data. This\nallows us to propose an effective constraint programming model to solve this\nproblem. We also discuss the model design and the different implementations of\nthe propagators that we used in practice. Finally, we show how this model,\ncoupled with a large neighborhood search, was able to find 24 hour schedules\nefficiently.\n", "date": "2023-12-21", "categories": []} {"id": "2312.10589", "title": "NN-Steiner: A Mixed Neural-algorithmic Approach for the Rectilinear\n Steiner Minimum Tree Problem", "abstract": " Recent years have witnessed rapid advances in the use of neural networks to\nsolve combinatorial optimization problems. Nevertheless, designing the \"right\"\nneural model that can effectively handle a given optimization problem can be\nchallenging, and often there is no theoretical understanding or justification\nof the resulting neural model. In this paper, we focus on the rectilinear\nSteiner minimum tree (RSMT) problem, which is of critical importance in IC\nlayout design and as a result has attracted numerous heuristic approaches in\nthe VLSI literature. Our contributions are two-fold. On the methodology front,\nwe propose NN-Steiner, which is a novel mixed neural-algorithmic framework for\ncomputing RSMTs that leverages the celebrated PTAS algorithmic framework of\nArora to solve this problem (and other geometric optimization problems). Our\nNN-Steiner replaces key algorithmic components within Arora's PTAS by suitable\nneural components. In particular, NN-Steiner only needs four neural network\n(NN) components that are called repeatedly within an algorithmic framework.\nCrucially, each of the four NN components is only of bounded size independent\nof input size, and thus easy to train. Furthermore, as the NN component is\nlearning a generic algorithmic step, once learned, the resulting mixed\nneural-algorithmic framework generalizes to much larger instances not seen in\ntraining. Our NN-Steiner, to our best knowledge, is the first neural\narchitecture of bounded size that has capacity to approximately solve RSMT (and\nvariants). On the empirical front, we show how NN-Steiner can be implemented\nand demonstrate the effectiveness of our resulting approach, especially in\nterms of generalization, by comparing with state-of-the-art methods (both\nneural and non-neural based).\n", "date": "2023-12-17", "categories": []} {"id": "2312.12554", "title": "Rectangle Search: An Anytime Beam Search (Extended Version)", "abstract": " Anytime heuristic search algorithms try to find a (potentially suboptimal)\nsolution as quickly as possible and then work to find better and better\nsolutions until an optimal solution is obtained or time is exhausted. The most\nwidely-known anytime search algorithms are based on best-first search. In this\npaper, we propose a new algorithm, rectangle search, that is instead based on\nbeam search, a variant of breadth-first search. It repeatedly explores\nalternatives at all depth levels and is thus best-suited to problems featuring\ndeep local minima. Experiments using a variety of popular search benchmarks\nsuggest that rectangle search is competitive with fixed-width beam search and\noften performs better than the previous best anytime search algorithms.\n", "date": "2023-12-19", "categories": []} {"id": "2312.11547", "title": "A Unified Pre-training and Adaptation Framework for Combinatorial\n Optimization on Graphs", "abstract": " Combinatorial optimization (CO) on graphs is a classic topic that has been\nextensively studied across many scientific and industrial fields. Recently,\nsolving CO problems on graphs through learning methods has attracted great\nattention. Advanced deep learning methods, e.g., graph neural networks (GNNs),\nhave been used to effectively assist the process of solving COs. However,\ncurrent frameworks based on GNNs are mainly designed for certain CO problems,\nthereby failing to consider their transferable and generalizable abilities\namong different COs on graphs. Moreover, simply using original graphs to model\nCOs only captures the direct correlations among objects, which does not\nconsider the mathematical logicality and properties of COs. In this paper, we\npropose a unified pre-training and adaptation framework for COs on graphs with\nthe help of the maximum satisfiability (Max-SAT) problem. We first use Max-SAT\nto bridge different COs on graphs since they can be converted to Max-SAT\nproblems represented by standard formulas and clauses with logical information.\nThen, we further design a pre-training and domain adaptation framework to\nextract the transferable and generalizable features so that different COs can\nbenefit from them. In the pre-training stage, Max-SAT instances are generated\nto initialize the parameters of the model. In the fine-tuning stage, instances\nfrom CO and Max-SAT problems are used for adaptation so that the transferable\nability can be further improved. Numerical experiments on several datasets show\nthat features extracted by our framework exhibit superior transferability and\nMax-SAT can boost the ability to solve COs on graphs.\n", "date": "2023-12-16", "categories": []} {"id": "2302.13036", "title": "Limited Query Graph Connectivity Test", "abstract": " We propose a combinatorial optimisation model called Limited Query Graph\nConnectivity Test. We consider a graph whose edges have two possible states\n(On/Off). The edges' states are hidden initially. We could query an edge to\nreveal its state. Given a source s and a destination t, we aim to test s-t\nconnectivity by identifying either a path (consisting of only On edges) or a\ncut (consisting of only Off edges). We are limited to B queries, after which we\nstop regardless of whether graph connectivity is established. We aim to design\na query policy that minimizes the expected number of queries.\n Our model is mainly motivated by a cyber security use case where we need to\nestablish whether an attack path exists in a network, between a source and a\ndestination. Edge query is resolved by manual effort from the IT admin, which\nis the motivation behind query minimization.\n Our model is highly related to monotone Stochastic Boolean Function\nEvaluation (SBFE). There are two existing exact algorithms for SBFE that are\nprohibitively expensive. We propose a significantly more scalable exact\nalgorithm. While previous exact algorithms only scale for trivial graphs (i.e.,\npast works experimented on at most 20 edges), we empirically demonstrate that\nour algorithm is scalable for a wide range of much larger practical graphs\n(i.e., Windows domain network graphs with tens of thousands of edges).\n We propose three heuristics. Our best-performing heuristic is via reducing\nthe search horizon of the exact algorithm. The other two are via reinforcement\nlearning (RL) and Monte Carlo tree search (MCTS). We also derive an anytime\nalgorithm for computing the performance lower bound. Experimentally, we show\nthat all our heuristics are near optimal. The exact algorithm based heuristic\noutperforms all, surpassing RL, MCTS and 8 existing heuristics ported from SBFE\nand related literature.\n", "date": "2023-02-25", "categories": []} {"id": "2312.10809", "title": "Deep-Dispatch: A Deep Reinforcement Learning-Based Vehicle Dispatch\n Algorithm for Advanced Air Mobility", "abstract": " Near future air taxi operations with electric vertical take-off and landing\n(eVTOL) aircraft will be constrained by the need for frequent recharging of\neVTOLs, limited takeoff and landing pads in vertiports, and subject to\ntime-varying demand and electricity prices, making the eVTOL dispatch problem\nunique and particularly challenging to solve. Previously, we have developed\noptimization models to address this problem. Such optimization models however\nsuffer from prohibitively high computational run times when the scale of the\nproblem increases, making them less practical for real world implementation. To\novercome this issue, we have developed two deep reinforcement learning-based\neVTOL dispatch algorithms, namely single-agent and multi-agent deep Q-learning\neVTOL dispatch algorithms, where the objective is to maximize operating profit.\nAn eVTOL-based passenger transportation simulation environment was built to\nassess the performance of our algorithms across $36$ numerical cases with\nvarying number of eVTOLs, vertiports, and demand. The results indicate that the\nmulti-agent eVTOL dispatch algorithm can closely approximate the optimal\ndispatch policy with significantly less computational expenses compared to the\nbenchmark optimization model. The multi-agent algorithm was found to outperform\nthe single-agent counterpart with respect to both profits generated and\ntraining time.\n", "date": "2023-12-17", "categories": []} {"id": "2312.10887", "title": "On Computing Makespan-Optimal Solutions for Generalized Sliding-Tile\n Puzzles", "abstract": " In the $15$-puzzle game, $15$ labeled square tiles are reconfigured on a\n$4\\times 4$ board through an escort, wherein each (time) step, a single tile\nneighboring it may slide into it, leaving the space previously occupied by the\ntile as the new escort. We study a generalized sliding-tile puzzle (GSTP) in\nwhich (1) there are $1+$ escorts and (2) multiple tiles can move synchronously\nin a single time step. Compared with popular discrete multi-agent/robot motion\nmodels, GSTP provides a more accurate model for a broad array of high-utility\napplications, including warehouse automation and autonomous garage parking, but\nis less studied due to the more involved tile interactions. In this work, we\nanalyze optimal GSTP solution structures, establishing that computing\nmakespan-optimal solutions for GSTP is NP-complete and developing polynomial\ntime algorithms yielding makespans approximating the minimum with expected/high\nprobability constant factors, assuming randomized start and goal\nconfigurations.\n", "date": "2023-12-18", "categories": []} {"id": "2312.09646", "title": "Exact Algorithms and Lowerbounds for Multiagent Pathfinding: Power of\n Treelike Topology", "abstract": " In the Multiagent Path Finding problem (MAPF for short), we focus on\nefficiently finding non-colliding paths for a set of $k$ agents on a given\ngraph $G$, where each agent seeks a path from its source vertex to a target. An\nimportant measure of the quality of the solution is the length of the proposed\nschedule $\\ell$, that is, the length of a longest path (including the waiting\ntime). In this work, we propose a systematic study under the parameterized\ncomplexity framework. The hardness results we provide align with many\nheuristics used for this problem, whose running time could potentially be\nimproved based on our fixed-parameter tractability results.\n We show that MAPF is W[1]-hard with respect to $k$ (even if $k$ is combined\nwith the maximum degree of the input graph). The problem remains NP-hard in\nplanar graphs even if the maximum degree and the makespan$\\ell$ are fixed\nconstants. On the positive side, we show an FPT algorithm for $k+\\ell$.\n As we delve further, the structure of~$G$ comes into play. We give an FPT\nalgorithm for parameter $k$ plus the diameter of the graph~$G$. The MAPF\nproblem is W[1]-hard for cliquewidth of $G$ plus $\\ell$ while it is FPT for\ntreewidth of $G$ plus $\\ell$.\n", "date": "2023-12-15", "categories": []} {"id": "2210.17178", "title": "Learning to Optimize Permutation Flow Shop Scheduling via Graph-based\n Imitation Learning", "abstract": " The permutation flow shop scheduling (PFSS), aiming at finding the optimal\npermutation of jobs, is widely used in manufacturing systems. When solving\nlarge-scale PFSS problems, traditional optimization algorithms such as\nheuristics could hardly meet the demands of both solution accuracy and\ncomputational efficiency, thus learning-based methods have recently garnered\nmore attention. Some work attempts to solve the problems by reinforcement\nlearning methods, which suffer from slow convergence issues during training and\nare still not accurate enough regarding the solutions. To that end, we propose\nto train the model via expert-driven imitation learning, which accelerates\nconvergence more stably and accurately. Moreover, in order to extract better\nfeature representations of input jobs, we incorporate the graph structure as\nthe encoder. The extensive experiments reveal that our proposed model obtains\nsignificant promotion and presents excellent generalizability in large-scale\nproblems with up to 1000 jobs. Compared to the state-of-the-art reinforcement\nlearning method, our model's network parameters are reduced to only 37\\% of\ntheirs, and the solution gap of our model towards the expert solutions\ndecreases from 6.8\\% to 1.3\\% on average. The code is available at:\n\\url{https://github.com/longkangli/PFSS-IL}.\n", "date": "2022-10-31", "categories": []} {"id": "2312.08103", "title": "Machine Learning for the Multi-Dimensional Bin Packing Problem:\n Literature Review and Empirical Evaluation", "abstract": " The Bin Packing Problem (BPP) is a well-established combinatorial\noptimization (CO) problem. Since it has many applications in our daily life,\ne.g. logistics and resource allocation, people are seeking efficient bin\npacking algorithms. On the other hand, researchers have been making constant\nadvances in machine learning (ML), which is famous for its efficiency. In this\narticle, we first formulate BPP, introducing its variants and practical\nconstraints. Then, a comprehensive survey on ML for multi-dimensional BPP is\nprovided. We further collect some public benchmarks of 3D BPP, and evaluate\nsome online methods on the Cutting Stock Dataset. Finally, we share our\nperspective on challenges and future directions in BPP. To the best of our\nknowledge, this is the first systematic review of ML-related methods for BPP.\n", "date": "2023-12-13", "categories": []} {"id": "2308.08641", "title": "Non-monotone Sequential Submodular Maximization", "abstract": " In this paper, we study a fundamental problem in submodular optimization,\nwhich is called sequential submodular maximization. Specifically, we aim to\nselect and rank a group of $k$ items from a ground set $V$ such that the\nweighted summation of $k$ (possibly non-monotone) submodular functions $f_1,\n\\cdots ,f_k: 2^V \\rightarrow \\mathbb{R}^+$ is maximized, here each function\n$f_j$ takes the first $j$ items from this sequence as input. The existing\nresearch on sequential submodular maximization has predominantly concentrated\non the monotone setting, assuming that the submodular functions are\nnon-decreasing. However, in various real-world scenarios, like diversity-aware\nrecommendation systems, adding items to an existing set might negatively impact\nthe overall utility. In response, this paper pioneers the examination of the\naforementioned problem with non-monotone submodular functions and offers\neffective solutions for both flexible and fixed length constraints, as well as\na special case with identical utility functions. The empirical evaluations\nfurther validate the effectiveness of our proposed algorithms in the domain of\nvideo recommendations. The results of this research have implications in\nvarious fields, including recommendation systems and assortment optimization,\nwhere the ordering of items significantly impacts the overall value obtained.\n", "date": "2023-08-16", "categories": []} {"id": "2311.16277", "title": "A Graph Neural Network-Based QUBO-Formulated Hamiltonian-Inspired Loss\n Function for Combinatorial Optimization using Reinforcement Learning", "abstract": " Quadratic Unconstrained Binary Optimization (QUBO) is a generic technique to\nmodel various NP-hard Combinatorial Optimization problems (CO) in the form of\nbinary variables. Ising Hamiltonian is used to model the energy function of a\nsystem. QUBO to Ising Hamiltonian is regarded as a technique to solve various\ncanonical optimization problems through quantum optimization algorithms.\nRecently, PI-GNN, a generic framework, has been proposed to address CO problems\nover graphs based on Graph Neural Network (GNN) architecture. They introduced a\ngeneric QUBO-formulated Hamiltonian-inspired loss function that was directly\noptimized using GNN. PI-GNN is highly scalable but there lies a noticeable\ndecrease in the number of satisfied constraints when compared to\nproblem-specific algorithms and becomes more pronounced with increased graph\ndensities. Here, We identify a behavioral pattern related to it and devise\nstrategies to improve its performance. Another group of literature uses\nReinforcement learning (RL) to solve the aforementioned NP-hard problems using\nproblem-specific reward functions. In this work, we also focus on creating a\nbridge between the RL-based solutions and the QUBO-formulated Hamiltonian. We\nformulate and empirically evaluate the compatibility of the QUBO-formulated\nHamiltonian as the generic reward function in the RL-based paradigm in the form\nof rewards. Furthermore, we also introduce a novel Monty Carlo Tree\nSearch-based strategy with GNN where we apply a guided search through manual\nperturbation of node labels during training. We empirically evaluated our\nmethods and observed up to 44% improvement in the number of constraint\nviolations compared to the PI-GNN.\n", "date": "2023-11-27", "categories": []} {"id": "2312.06877", "title": "A Novel Differentiable Loss Function for Unsupervised Graph Neural\n Networks in Graph Partitioning", "abstract": " In this paper, we explore the graph partitioning problem, a pivotal\ncombina-torial optimization challenge with extensive applications in various\nfields such as science, technology, and business. Recognized as an NP-hard\nprob-lem, graph partitioning lacks polynomial-time algorithms for its\nresolution. Recently, there has been a burgeoning interest in leveraging\nmachine learn-ing, particularly approaches like supervised, unsupervised, and\nreinforce-ment learning, to tackle such NP-hard problems. However, these\nmethods face significant hurdles: supervised learning is constrained by the\nnecessity of labeled solution instances, which are often computationally\nimpractical to obtain; reinforcement learning grapples with instability in the\nlearning pro-cess; and unsupervised learning contends with the absence of a\ndifferentia-ble loss function, a consequence of the discrete nature of most\ncombinatorial optimization problems. Addressing these challenges, our research\nintroduces a novel pipeline employing an unsupervised graph neural network to\nsolve the graph partitioning problem. The core innovation of this study is the\nfor-mulation of a differentiable loss function tailored for this purpose. We\nrigor-ously evaluate our methodology against contemporary state-of-the-art\ntech-niques, focusing on metrics: cuts and balance, and our findings reveal\nthat our is competitive with these leading methods.\n", "date": "2023-12-11", "categories": []} {"id": "2211.13436", "title": "Solving Bilevel Knapsack Problem using Graph Neural Networks", "abstract": " The Bilevel Optimization Problem is a hierarchical optimization problem with\ntwo agents, a leader and a follower. The leader make their own decisions first,\nand the followers make the best choices accordingly. The leader knows the\ninformation of the followers, and the goal of the problem is to find the\noptimal solution by considering the reactions of the followers from the\nleader's point of view. For the Bilevel Optimization Problem, there are no\ngeneral and efficient algorithms or commercial solvers to get an optimal\nsolution, and it is very difficult to get a good solution even for a simple\nproblem. In this paper, we propose a deep learning approach using Graph Neural\nNetworks to solve the bilevel knapsack problem. We train the model to predict\nthe leader's solution and use it to transform the hierarchical optimization\nproblem into a single-level optimization problem to get the solution. Our model\nfound the feasible solution that was about 500 times faster than the exact\nalgorithm with $1.7\\%$ optimal gap. Also, our model performed well on problems\nof different size from the size it was trained on.\n", "date": "2022-11-24", "categories": []} {"id": "2303.06024", "title": "A hybrid deep-learning-metaheuristic framework for bi-level network\n design problems", "abstract": " This study proposes a hybrid deep-learning-metaheuristic framework with a\nbi-level architecture for road network design problems (NDPs). We train a graph\nneural network (GNN) to approximate the solution of the user equilibrium (UE)\ntraffic assignment problem and use inferences made by the trained model to\ncalculate fitness function evaluations of a genetic algorithm (GA) to\napproximate solutions for NDPs. Using three test networks, two NDP variants and\nan exact solver as benchmark, we show that on average, our proposed framework\ncan provide solutions within 1.5% gap of the best results in less than 0.5% of\nthe time used by the exact solution procedure. Our framework can be utilized\nwithin an expert system for infrastructure planning to determine the best\ninfrastructure planning and management decisions under different scenarios.\nGiven the flexibility of the framework, it can easily be adapted to many other\ndecision problems that can be modeled as bi-level problems on graphs. Moreover,\nwe foreseen interesting future research directions, thus we also put forward a\nbrief research agenda for this topic. The key observation from our research\nthat can shape future research is that the fitness function evaluation time\nusing the inferences made by the GNN model was in the order of milliseconds,\nwhich points to an opportunity and a need for novel heuristics that 1) can cope\nwell with noisy fitness function values provided by deep learning models, and\n2) can use the significantly enlarged efficiency of the evaluation step to\nexplore the search space effectively (rather than efficiently). This opens a\nnew avenue for a modern class of metaheuristics that are crafted for use with\nAI-powered predictors.\n", "date": "2023-03-10", "categories": []} {"id": "2306.09803", "title": "Framework and Benchmarks for Combinatorial and Mixed-variable Bayesian\n Optimization", "abstract": " This paper introduces a modular framework for Mixed-variable and\nCombinatorial Bayesian Optimization (MCBO) to address the lack of systematic\nbenchmarking and standardized evaluation in the field. Current MCBO papers\noften introduce non-diverse or non-standard benchmarks to evaluate their\nmethods, impeding the proper assessment of different MCBO primitives and their\ncombinations. Additionally, papers introducing a solution for a single MCBO\nprimitive often omit benchmarking against baselines that utilize the same\nmethods for the remaining primitives. This omission is primarily due to the\nsignificant implementation overhead involved, resulting in a lack of controlled\nassessments and an inability to showcase the merits of a contribution\neffectively. To overcome these challenges, our proposed framework enables an\neffortless combination of Bayesian Optimization components, and provides a\ndiverse set of synthetic and real-world benchmarking tasks. Leveraging this\nflexibility, we implement 47 novel MCBO algorithms and benchmark them against\nseven existing MCBO solvers and five standard black-box optimization algorithms\non ten tasks, conducting over 4000 experiments. Our findings reveal a superior\ncombination of MCBO primitives outperforming existing approaches and illustrate\nthe significance of model fit and the use of a trust region. We make our MCBO\nlibrary available under the MIT license at\n\\url{https://github.com/huawei-noah/HEBO/tree/master/MCBO}.\n", "date": "2023-06-16", "categories": []} {"id": "2312.05980", "title": "Maximum flow-based formulation for the optimal location of electric\n vehicle charging stations", "abstract": " With the increasing effects of climate change, the urgency to step away from\nfossil fuels is greater than ever before. Electric vehicles (EVs) are one way\nto diminish these effects, but their widespread adoption is often limited by\nthe insufficient availability of charging stations. In this work, our goal is\nto expand the infrastructure of EV charging stations, in order to provide a\nbetter quality of service in terms of user satisfaction (and availability of\ncharging stations). Specifically, our focus is directed towards urban areas. We\nfirst propose a model for the assignment of EV charging demand to stations,\nframing it as a maximum flow problem. This model is the basis for the\nevaluation of user satisfaction with a given charging infrastructure. Secondly,\nwe incorporate the maximum flow model into a mixed-integer linear program,\nwhere decisions on the opening of new stations and on the expansion of their\ncapacity through additional outlets is accounted for. We showcase our\nmethodology for the city of Montreal, demonstrating the scalability of our\napproach to handle real-world scenarios. We conclude that considering both\nspacial and temporal variations in charging demand is meaningful when solving\nrealistic instances.\n", "date": "2023-12-10", "categories": []} {"id": "2312.03413", "title": "Approximating Solutions to the Knapsack Problem using the Lagrangian\n Dual Framework", "abstract": " The Knapsack Problem is a classic problem in combinatorial optimisation.\nSolving these problems may be computationally expensive. Recent years have seen\na growing interest in the use of deep learning methods to approximate the\nsolutions to such problems. A core problem is how to enforce or encourage\nconstraint satisfaction in predicted solutions. A promising approach for\npredicting solutions to constrained optimisation problems is the Lagrangian\nDual Framework which builds on the method of Lagrangian Relaxation. In this\npaper we develop neural network models to approximate Knapsack Problem\nsolutions using the Lagrangian Dual Framework while improving constraint\nsatisfaction. We explore the problems of output interpretation and model\nselection within this context. Experimental results show strong constraint\nsatisfaction with a minor reduction of optimality as compared to a baseline\nneural network which does not explicitly model the constraints.\n", "date": "2023-12-06", "categories": []} {"id": "2312.02720", "title": "Towards the Inferrence of Structural Similarity of Combinatorial\n Landscapes", "abstract": " One of the most common problem-solving heuristics is by analogy. For a given\nproblem, a solver can be viewed as a strategic walk on its fitness landscape.\nThus if a solver works for one problem instance, we expect it will also be\neffective for other instances whose fitness landscapes essentially share\nstructural similarities with each other. However, due to the black-box nature\nof combinatorial optimization, it is far from trivial to infer such similarity\nin real-world scenarios. To bridge this gap, by using local optima network as a\nproxy of fitness landscapes, this paper proposed to leverage graph data mining\ntechniques to conduct qualitative and quantitative analyses to explore the\nlatent topological structural information embedded in those landscapes. By\nconducting large-scale empirical experiments on three classic combinatorial\noptimization problems, we gain concrete evidence to support the existence of\nstructural similarity between landscapes of the same classes within neighboring\ndimensions. We also interrogated the relationship between landscapes of\ndifferent problem classes.\n", "date": "2023-12-05", "categories": []} {"id": "2302.08224", "title": "DIFUSCO: Graph-based Diffusion Solvers for Combinatorial Optimization", "abstract": " Neural network-based Combinatorial Optimization (CO) methods have shown\npromising results in solving various NP-complete (NPC) problems without relying\non hand-crafted domain knowledge. This paper broadens the current scope of\nneural solvers for NPC problems by introducing a new graph-based diffusion\nframework, namely DIFUSCO. Our framework casts NPC problems as discrete {0,\n1}-vector optimization problems and leverages graph-based denoising diffusion\nmodels to generate high-quality solutions. We investigate two types of\ndiffusion models with Gaussian and Bernoulli noise, respectively, and devise an\neffective inference schedule to enhance the solution quality. We evaluate our\nmethods on two well-studied NPC combinatorial optimization problems: Traveling\nSalesman Problem (TSP) and Maximal Independent Set (MIS). Experimental results\nshow that DIFUSCO strongly outperforms the previous state-of-the-art neural\nsolvers, improving the performance gap between ground-truth and neural solvers\nfrom 1.76% to 0.46% on TSP-500, from 2.46% to 1.17% on TSP-1000, and from 3.19%\nto 2.58% on TSP10000. For the MIS problem, DIFUSCO outperforms the previous\nstate-of-the-art neural solver on the challenging SATLIB benchmark.\n", "date": "2023-02-16", "categories": []} {"id": "2312.00140", "title": "The Stochastic Dynamic Post-Disaster Inventory Allocation Problem with\n Trucks and UAVs", "abstract": " Humanitarian logistics operations face increasing difficulties due to rising\ndemands for aid in disaster areas. This paper investigates the dynamic\nallocation of scarce relief supplies across multiple affected districts over\ntime. It introduces a novel stochastic dynamic post-disaster inventory\nallocation problem with trucks and unmanned aerial vehicles delivering relief\ngoods under uncertain supply and demand. The relevance of this humanitarian\nlogistics problem lies in the importance of considering the inter-temporal\nsocial impact of deliveries. We achieve this by incorporating deprivation costs\nwhen allocating scarce supplies. Furthermore, we consider the inherent\nuncertainties of disaster areas and the potential use of cargo UAVs to enhance\noperational efficiency. This study proposes two anticipatory solution methods\nbased on approximate dynamic programming, specifically decomposed linear value\nfunction approximation and neural network value function approximation to\neffectively manage uncertainties in the dynamic allocation process. We compare\nDL-VFA and NN-VFA with various state-of-the-art methods (exact re-optimization,\nPPO) and results show a 6-8% improvement compared to the best benchmarks.\nNN-VFA provides the best performance and captures nonlinearities in the\nproblem, whereas DL-VFA shows excellent scalability against a minor performance\nloss. The experiments reveal that consideration of deprivation costs results in\nimproved allocation of scarce supplies both across affected districts and over\ntime. Finally, results show that deploying UAVs can play a crucial role in the\nallocation of relief goods, especially in the first stages after a disaster.\nThe use of UAVs reduces transportation- and deprivation costs together by\n16-20% and reduces maximum deprivation times by 19-40%, while maintaining\nsimilar levels of demand coverage, showcasing efficient and effective\noperations.\n", "date": "2023-11-30", "categories": []} {"id": "2311.17393", "title": "Comparison of metaheuristics for the firebreak placement problem: a\n simulation-based optimization approach", "abstract": " The problem of firebreak placement is crucial for fire prevention, and its\neffectiveness at landscape scale will depend on their ability to impede the\nprogress of future wildfires. To provide an adequate response, it is therefore\nnecessary to consider the stochastic nature of fires, which are highly\nunpredictable from ignition to extinction. Thus, the placement of firebreaks\ncan be considered a stochastic optimization problem where: (1) the objective\nfunction is to minimize the expected cells burnt of the landscape; (2) the\ndecision variables being the location of firebreaks; and (3) the random\nvariable being the spatial propagation/behavior of fires. In this paper, we\npropose a solution approach for the problem from the perspective of\nsimulation-based optimization (SbO), where the objective function is not\navailable (a black-box function), but can be computed (and/or approximated) by\nwildfire simulations. For this purpose, Genetic Algorithm and GRASP are\nimplemented. The final implementation yielded favorable results for the Genetic\nAlgorithm, demonstrating strong performance in scenarios with medium to high\noperational capacity, as well as medium levels of stochasticity\n", "date": "2023-11-29", "categories": []} {"id": "2311.17766", "title": "Robustness Approaches for the Examination Timetabling Problem under Data\n Uncertainty", "abstract": " In the literature the examination timetabling problem (ETTP) is often\nconsidered a post-enrollment problem (PE-ETTP). In the real world, universities\noften schedule their exams before students register using information from\nprevious terms. A direct consequence of this approach is the uncertainty\npresent in the resulting models. In this work we discuss several approaches\navailable in the robust optimization literature. We consider the implications\nof each approach in respect to the examination timetabling problem and present\nhow the most favorable approaches can be applied to the ETTP. Afterwards we\nanalyze the impact of some possible implementations of the given robustness\napproaches on two real world instances and several random instances generated\nby our instance generation framework which we introduce in this work.\n", "date": "2023-11-29", "categories": []} {"id": "2311.16171", "title": "Multi-Agent Learning of Efficient Fulfilment and Routing Strategies in\n E-Commerce", "abstract": " This paper presents an integrated algorithmic framework for minimising\nproduct delivery costs in e-commerce (known as the cost-to-serve or C2S). One\nof the major challenges in e-commerce is the large volume of spatio-temporally\ndiverse orders from multiple customers, each of which has to be fulfilled from\none of several warehouses using a fleet of vehicles. This results in two levels\nof decision-making: (i) selection of a fulfillment node for each order\n(including the option of deferral to a future time), and then (ii) routing of\nvehicles (each of which can carry multiple orders originating from the same\nwarehouse). We propose an approach that combines graph neural networks and\nreinforcement learning to train the node selection and vehicle routing agents.\nWe include real-world constraints such as warehouse inventory capacity, vehicle\ncharacteristics such as travel times, service times, carrying capacity, and\ncustomer constraints including time windows for delivery. The complexity of\nthis problem arises from the fact that outcomes (rewards) are driven both by\nthe fulfillment node mapping as well as the routing algorithms, and are\nspatio-temporally distributed. Our experiments show that this algorithmic\npipeline outperforms pure heuristic policies.\n", "date": "2023-11-20", "categories": []} {"id": "2311.15216", "title": "Solve Large-scale Unit Commitment Problems by Physics-informed Graph\n Learning", "abstract": " Unit commitment (UC) problems are typically formulated as mixed-integer\nprograms (MIP) and solved by the branch-and-bound (B&B) scheme. The recent\nadvances in graph neural networks (GNN) enable it to enhance the B&B algorithm\nin modern MIP solvers by learning to dive and branch. Existing GNN models that\ntackle MIP problems are mostly constructed from mathematical formulation, which\nis computationally expensive when dealing with large-scale UC problems. In this\npaper, we propose a physics-informed hierarchical graph convolutional network\n(PI-GCN) for neural diving that leverages the underlying features of various\ncomponents of power systems to find high-quality variable assignments.\nFurthermore, we adopt the MIP model-based graph convolutional network (MB-GCN)\nfor neural branching to select the optimal variables for branching at each node\nof the B&B tree. Finally, we integrate neural diving and neural branching into\na modern MIP solver to establish a novel neural MIP solver designed for\nlarge-scale UC problems. Numeral studies show that PI-GCN has better\nperformance and scalability than the baseline MB-GCN on neural diving.\nMoreover, the neural MIP solver yields the lowest operational cost and\noutperforms a modern MIP solver for all testing days after combining it with\nour proposed neural diving model and the baseline neural branching model.\n", "date": "2023-11-26", "categories": []} {"id": "2311.16030", "title": "Machine Learning-Enhanced Aircraft Landing Scheduling under\n Uncertainties", "abstract": " This paper addresses aircraft delays, emphasizing their impact on safety and\nfinancial losses. To mitigate these issues, an innovative machine learning\n(ML)-enhanced landing scheduling methodology is proposed, aiming to improve\nautomation and safety. Analyzing flight arrival delay scenarios reveals strong\nmultimodal distributions and clusters in arrival flight time durations. A\nmulti-stage conditional ML predictor enhances separation time prediction based\non flight events. ML predictions are then integrated as safety constraints in a\ntime-constrained traveling salesman problem formulation, solved using\nmixed-integer linear programming (MILP). Historical flight recordings and model\npredictions address uncertainties between successive flights, ensuring\nreliability. The proposed method is validated using real-world data from the\nAtlanta Air Route Traffic Control Center (ARTCC ZTL). Case studies demonstrate\nan average 17.2% reduction in total landing time compared to the\nFirst-Come-First-Served (FCFS) rule. Unlike FCFS, the proposed methodology\nconsiders uncertainties, instilling confidence in scheduling. The study\nconcludes with remarks and outlines future research directions.\n", "date": "2023-11-27", "categories": []} {"id": "2311.13695", "title": "BackboneLearn: A Library for Scaling Mixed-Integer Optimization-Based\n Machine Learning", "abstract": " We present BackboneLearn: an open-source software package and framework for\nscaling mixed-integer optimization (MIO) problems with indicator variables to\nhigh-dimensional problems. This optimization paradigm can naturally be used to\nformulate fundamental problems in interpretable supervised learning (e.g.,\nsparse regression and decision trees), in unsupervised learning (e.g.,\nclustering), and beyond; BackboneLearn solves the aforementioned problems\nfaster than exact methods and with higher accuracy than commonly used\nheuristics. The package is built in Python and is user-friendly and easily\nextensible: users can directly implement a backbone algorithm for their MIO\nproblem at hand. The source code of BackboneLearn is available on GitHub (link:\nhttps://github.com/chziakas/backbone_learn).\n", "date": "2023-11-22", "categories": []} {"id": "2311.13843", "title": "Exact Combinatorial Optimization with Temporo-Attentional Graph Neural\n Networks", "abstract": " Combinatorial optimization finds an optimal solution within a discrete set of\nvariables and constraints. The field has seen tremendous progress both in\nresearch and industry. With the success of deep learning in the past decade, a\nrecent trend in combinatorial optimization has been to improve state-of-the-art\ncombinatorial optimization solvers by replacing key heuristic components with\nmachine learning (ML) models. In this paper, we investigate two essential\naspects of machine learning algorithms for combinatorial optimization: temporal\ncharacteristics and attention. We argue that for the task of variable selection\nin the branch-and-bound (B&B) algorithm, incorporating the temporal information\nas well as the bipartite graph attention improves the solver's performance. We\nsupport our claims with intuitions and numerical results over several standard\ndatasets used in the literature and competitions. Code is available at:\nhttps://developer.huaweicloud.com/develop/aigallery/notebook/detail?id=047c6cf2-8463-40d7-b92f-7b2ca998e935\n", "date": "2023-11-23", "categories": []} {"id": "2311.14156", "title": "Variational Annealing on Graphs for Combinatorial Optimization", "abstract": " Several recent unsupervised learning methods use probabilistic approaches to\nsolve combinatorial optimization (CO) problems based on the assumption of\nstatistically independent solution variables. We demonstrate that this\nassumption imposes performance limitations in particular on difficult problem\ninstances. Our results corroborate that an autoregressive approach which\ncaptures statistical dependencies among solution variables yields superior\nperformance on many popular CO problems. We introduce subgraph tokenization in\nwhich the configuration of a set of solution variables is represented by a\nsingle token. This tokenization technique alleviates the drawback of the long\nsequential sampling procedure which is inherent to autoregressive methods\nwithout sacrificing expressivity. Importantly, we theoretically motivate an\nannealed entropy regularization and show empirically that it is essential for\nefficient and stable learning.\n", "date": "2023-11-23", "categories": []} {"id": "2311.14595", "title": "A Survey and Analysis of Evolutionary Operators for Permutations", "abstract": " There are many combinatorial optimization problems whose solutions are best\nrepresented by permutations. The classic traveling salesperson seeks an optimal\nordering over a set of cities. Scheduling problems often seek optimal orderings\nof tasks or activities. Although some evolutionary approaches to such problems\nutilize the bit strings of a genetic algorithm, it is more common to directly\nrepresent solutions with permutations. Evolving permutations directly requires\nspecialized evolutionary operators. Over the years, many crossover and mutation\noperators have been developed for solving permutation problems with\nevolutionary algorithms. In this paper, we survey the breadth of evolutionary\noperators for permutations. We implemented all of these in Chips-n-Salsa, an\nopen source Java library for evolutionary computation. Finally, we empirically\nanalyze the crossover operators on artificial fitness landscapes isolating\ndifferent permutation features.\n", "date": "2023-11-24", "categories": []} {"id": "2311.12802", "title": "A general Framework for Utilizing Metaheuristic Optimization for\n Sustainable Unrelated Parallel Machine Scheduling: A concise overview", "abstract": " Sustainable development has emerged as a global priority, and industries are\nincreasingly striving to align their operations with sustainable practices.\nParallel machine scheduling (PMS) is a critical aspect of production planning\nthat directly impacts resource utilization and operational efficiency. In this\npaper, we investigate the application of metaheuristic optimization algorithms\nto address the unrelated parallel machine scheduling problem (UPMSP) through\nthe lens of sustainable development goals (SDGs). The primary objective of this\nstudy is to explore how metaheuristic optimization algorithms can contribute to\nachieving sustainable development goals in the context of UPMSP. We examine a\nrange of metaheuristic algorithms, including genetic algorithms, particle swarm\noptimization, ant colony optimization, and more, and assess their effectiveness\nin optimizing the scheduling problem. The algorithms are evaluated based on\ntheir ability to improve resource utilization, minimize energy consumption,\nreduce environmental impact, and promote socially responsible production\npractices. To conduct a comprehensive analysis, we consider UPMSP instances\nthat incorporate sustainability-related constraints and objectives.\n", "date": "2023-09-14", "categories": []} {"id": "2305.17010", "title": "Let the Flows Tell: Solving Graph Combinatorial Optimization Problems\n with GFlowNets", "abstract": " Combinatorial optimization (CO) problems are often NP-hard and thus out of\nreach for exact algorithms, making them a tempting domain to apply machine\nlearning methods. The highly structured constraints in these problems can\nhinder either optimization or sampling directly in the solution space. On the\nother hand, GFlowNets have recently emerged as a powerful machinery to\nefficiently sample from composite unnormalized densities sequentially and have\nthe potential to amortize such solution-searching processes in CO, as well as\ngenerate diverse solution candidates. In this paper, we design Markov decision\nprocesses (MDPs) for different combinatorial problems and propose to train\nconditional GFlowNets to sample from the solution space. Efficient training\ntechniques are also developed to benefit long-range credit assignment. Through\nextensive experiments on a variety of different CO tasks with synthetic and\nrealistic data, we demonstrate that GFlowNet policies can efficiently find\nhigh-quality solutions. Our implementation is open-sourced at\nhttps://github.com/zdhNarsil/GFlowNet-CombOpt.\n", "date": "2023-05-26", "categories": []} {"id": "2310.12629", "title": "Online Combinatorial Linear Optimization via a Frank-Wolfe-based\n Metarounding Algorithm", "abstract": " Metarounding is an approach to convert an approximation algorithm for linear\noptimization over some combinatorial classes to an online linear optimization\nalgorithm for the same class. We propose a new metarounding algorithm under a\nnatural assumption that a relax-based approximation algorithm exists for the\ncombinatorial class. Our algorithm is much more efficient in both theoretical\nand practical aspects.\n", "date": "2023-10-19", "categories": []} {"id": "2310.15543", "title": "Symmetry-preserving graph attention network to solve routing problems at\n multiple resolutions", "abstract": " Travelling Salesperson Problems (TSPs) and Vehicle Routing Problems (VRPs)\nhave achieved reasonable improvement in accuracy and computation time with the\nadaptation of Machine Learning (ML) methods. However, none of the previous\nworks completely respects the symmetries arising from TSPs and VRPs including\nrotation, translation, permutation, and scaling. In this work, we introduce the\nfirst-ever completely equivariant model and training to solve combinatorial\nproblems. Furthermore, it is essential to capture the multiscale structure\n(i.e. from local to global information) of the input graph, especially for the\ncases of large and long-range graphs, while previous methods are limited to\nextracting only local information that can lead to a local or sub-optimal\nsolution. To tackle the above limitation, we propose a Multiresolution scheme\nin combination with Equivariant Graph Attention network (mEGAT) architecture,\nwhich can learn the optimal route based on low-level and high-level graph\nresolutions in an efficient way. In particular, our approach constructs a\nhierarchy of coarse-graining graphs from the input graph, in which we try to\nsolve the routing problems on simple low-level graphs first, then utilize that\nknowledge for the more complex high-level graphs. Experimentally, we have shown\nthat our model outperforms existing baselines and proved that symmetry\npreservation and multiresolution are important recipes for solving\ncombinatorial problems in a data-driven manner. Our source code is publicly\navailable at https://github.com/HySonLab/Multires-NP-hard\n", "date": "2023-10-24", "categories": []} {"id": "2311.11400", "title": "Make me an Offer: Forward and Reverse Auctioning Problems in the Tourism\n Industry", "abstract": " Most tourist destinations are facing regular and consistent seasonality with\nsignificant economic and social impacts. This phenomenon is more pronounced in\nthe post-covid era, where demand for travel has increased but unevenly among\ndifferent geographic areas. To counter these problems that both customers and\nhoteliers are facing, we have developed two auctioning systems that allow\nhoteliers of lower popularity tier areas or during low season periods to\nauction their rooms in what we call a forward auction model, and also allows\ncustomers to initiate a bidding process whereby hoteliers in an area may make\noffers to the customer for their rooms, in what constitutes a reverse auction\nmodel initiated by the customer, similar to the bidding concept of\npriceline.com. We develop mathematical programming models that define\nexplicitly both types of auctions, and show that in each type, there are\nsignificant benefits to be gained both on the side of the hotelier as well as\non the side of the customer. We discuss algorithmic techniques for the\napproximate solution of these optimization problems, and present results using\nexact optimization solvers to solve them to guaranteed optimality. These\ntechniques could be beneficial to both customer and hotelier reducing\nseasonality during middle and low season and providing the customer with\nattractive offers.\n", "date": "2023-11-19", "categories": []} {"id": "2311.10316", "title": "Graph Sparsifications using Neural Network Assisted Monte Carlo Tree\n Search", "abstract": " Graph neural networks have been successful for machine learning, as well as\nfor combinatorial and graph problems such as the Subgraph Isomorphism Problem\nand the Traveling Salesman Problem. We describe an approach for computing graph\nsparsifiers by combining a graph neural network and Monte Carlo Tree Search. We\nfirst train a graph neural network that takes as input a partial solution and\nproposes a new node to be added as output. This neural network is then used in\na Monte Carlo search to compute a sparsifier. The proposed method consistently\noutperforms several standard approximation algorithms on different types of\ngraphs and often finds the optimal solution.\n", "date": "2023-11-17", "categories": []} {"id": "2311.09275", "title": "Improved Sparse Ising Optimization", "abstract": " Sparse Ising problems can be found in application areas such as logistics,\ncondensed matter physics and training of deep Boltzmann networks, but can be\nvery difficult to tackle with high efficiency and accuracy. This report\npresents new data demonstrating significantly higher performance on some\nlongstanding benchmark problems with up to 20,000 variables. The data come from\na new heuristic algorithm tested on the large sparse instances from the Gset\nbenchmark suite. Relative to leading reported combinations of speed and\naccuracy (e.g., from Toshiba's Simulated Bifurcation Machine and Breakout Local\nSearch), a proof-of-concept implementation reached targets 2-4 orders of\nmagnitude faster. For two instances (G72 and G77) the new algorithm discovered\na better solution than all previously reported values. Solution bitstrings\nconfirming these two best solutions are provided. The data suggest exciting\npossibilities for pushing the sparse Ising performance frontier to potentially\nstrengthen algorithm portfolios, AI toolkits and decision-making systems.\n", "date": "2023-11-15", "categories": []} {"id": "2311.08834", "title": "A* search algorithm for an optimal investment problem in vehicle-sharing\n systems", "abstract": " We study an optimal investment problem that arises in the context of the\nvehicle-sharing system. Given a set of locations to build stations, we need to\ndetermine i) the sequence of stations to be built and the number of vehicles to\nacquire in order to obtain the target state where all stations are built, and\nii) the number of vehicles to acquire and their allocation in order to maximize\nthe total profit returned by operating the system when some or all stations are\nopen. The profitability associated with operating open stations, measured over\na specific time period, is represented as a linear optimization problem applied\nto a collection of open stations. With operating capital, the owner of the\nsystem can open new stations. This property introduces a set-dependent aspect\nto the duration required for opening a new station, and the optimal investment\nproblem can be viewed as a variant of the Traveling Salesman Problem (TSP) with\nset-dependent cost. We propose an A* search algorithm to address this\nparticular variant of the TSP. Computational experiments highlight the benefits\nof the proposed algorithm in comparison to the widely recognized Dijkstra\nalgorithm and propose future research to explore new possibilities and\napplications for both exact and approximate A* algorithms.\n", "date": "2023-11-15", "categories": []} {"id": "2306.13258", "title": "Fast Maximum $k$-Plex Algorithms Parameterized by Small Degeneracy Gaps", "abstract": " Given a graph, a $k$-plex is a set of vertices in which each vertex is not\nadjacent to at most $k-1$ other vertices in the set. The maximum $k$-plex\nproblem, which asks for the largest $k$-plex from the given graph, is an\nimportant but computationally challenging problem in applications such as graph\nmining and community detection. So far, there are many practical algorithms,\nbut without providing theoretical explanations on their efficiency. We define a\nnovel parameter of the input instance, $g_k(G)$, the gap between the degeneracy\nbound and the size of the maximum $k$-plex in the given graph, and present an\nexact algorithm parameterized by this $g_k(G)$, which has a worst-case running\ntime polynomial in the size of the input graph and exponential in $g_k(G)$. In\nreal-world inputs, $g_k(G)$ is very small, usually bounded by $O(\\log{(|V|)})$,\nindicating that the algorithm runs in polynomial time. We further extend our\ndiscussion to an even smaller parameter $cg_k(G)$, the gap between the\ncommunity-degeneracy bound and the size of the maximum $k$-plex, and show that\nwithout much modification, our algorithm can also be parameterized by\n$cg_k(G)$. To verify the empirical performance of these algorithms, we carry\nout extensive experiments to show that these algorithms are competitive with\nthe state-of-the-art algorithms. In particular, for large $k$ values such as\n$15$ and $20$, our algorithms dominate the existing algorithms. Finally,\nempirical analysis is performed to illustrate the effectiveness of the\nparameters and other key components in the implementation.\n", "date": "2023-06-23", "categories": []} {"id": "2311.07708", "title": "Reinforcement Learning for Solving Stochastic Vehicle Routing Problem", "abstract": " This study addresses a gap in the utilization of Reinforcement Learning (RL)\nand Machine Learning (ML) techniques in solving the Stochastic Vehicle Routing\nProblem (SVRP) that involves the challenging task of optimizing vehicle routes\nunder uncertain conditions. We propose a novel end-to-end framework that\ncomprehensively addresses the key sources of stochasticity in SVRP and utilizes\nan RL agent with a simple yet effective architecture and a tailored training\nmethod. Through comparative analysis, our proposed model demonstrates superior\nperformance compared to a widely adopted state-of-the-art metaheuristic,\nachieving a significant 3.43% reduction in travel costs. Furthermore, the model\nexhibits robustness across diverse SVRP settings, highlighting its adaptability\nand ability to learn optimal routing strategies in varying environments. The\npublicly available implementation of our framework serves as a valuable\nresource for future research endeavors aimed at advancing RL-based solutions\nfor SVRP.\n", "date": "2023-11-13", "categories": []} {"id": "2311.06972", "title": "An Expandable Machine Learning-Optimization Framework to Sequential\n Decision-Making", "abstract": " We present an integrated prediction-optimization (PredOpt) framework to\nefficiently solve sequential decision-making problems by predicting the values\nof binary decision variables in an optimal solution. We address the key issues\nof sequential dependence, infeasibility, and generalization in machine learning\n(ML) to make predictions for optimal solutions to combinatorial problems. The\nsequential nature of the combinatorial optimization problems considered is\ncaptured with recurrent neural networks and a sliding-attention window. We\nintegrate an attention-based encoder-decoder neural network architecture with\nan infeasibility-elimination and generalization framework to learn high-quality\nfeasible solutions to time-dependent optimization problems. In this framework,\nthe required level of predictions is optimized to eliminate the infeasibility\nof the ML predictions. These predictions are then fixed in mixed-integer\nprogramming (MIP) problems to solve them quickly with the aid of a commercial\nsolver. We demonstrate our approach to tackling the two well-known dynamic\nNP-Hard optimization problems: multi-item capacitated lot-sizing (MCLSP) and\nmulti-dimensional knapsack (MSMK). Our results show that models trained on\nshorter and smaller-dimensional instances can be successfully used to predict\nlonger and larger-dimensional problems. The solution time can be reduced by\nthree orders of magnitude with an average optimality gap below 0.1%. We compare\nPredOpt with various specially designed heuristics and show that our framework\noutperforms them. PredOpt can be advantageous for solving dynamic MIP problems\nthat need to be solved instantly and repetitively.\n", "date": "2023-11-12", "categories": []} {"id": "2311.05650", "title": "Learning to Configure Separators in Branch-and-Cut", "abstract": " Cutting planes are crucial in solving mixed integer linear programs (MILP) as\nthey facilitate bound improvements on the optimal solution. Modern MILP solvers\nrely on a variety of separators to generate a diverse set of cutting planes by\ninvoking the separators frequently during the solving process. This work\nidentifies that MILP solvers can be drastically accelerated by appropriately\nselecting separators to activate. As the combinatorial separator selection\nspace imposes challenges for machine learning, we learn to separate by\nproposing a novel data-driven strategy to restrict the selection space and a\nlearning-guided algorithm on the restricted space. Our method predicts\ninstance-aware separator configurations which can dynamically adapt during the\nsolve, effectively accelerating the open source MILP solver SCIP by improving\nthe relative solve time up to 72% and 37% on synthetic and real-world MILP\nbenchmarks. Our work complements recent work on learning to select cutting\nplanes and highlights the importance of separator management.\n", "date": "2023-11-08", "categories": []} {"id": "2311.06063", "title": "RIGA: A Regret-Based Interactive Genetic Algorithm", "abstract": " In this paper, we propose an interactive genetic algorithm for solving\nmulti-objective combinatorial optimization problems under preference\nimprecision. More precisely, we consider problems where the decision maker's\npreferences over solutions can be represented by a parameterized aggregation\nfunction (e.g., a weighted sum, an OWA operator, a Choquet integral), and we\nassume that the parameters are initially not known by the recommendation\nsystem. In order to quickly make a good recommendation, we combine elicitation\nand search in the following way: 1) we use regret-based elicitation techniques\nto reduce the parameter space in a efficient way, 2) genetic operators are\napplied on parameter instances (instead of solutions) to better explore the\nparameter space, and 3) we generate promising solutions (population) using\nexisting solving methods designed for the problem with known preferences. Our\nalgorithm, called RIGA, can be applied to any multi-objective combinatorial\noptimization problem provided that the aggregation function is linear in its\nparameters and that a (near-)optimal solution can be efficiently determined for\nthe problem with known preferences. We also study its theoretical performances:\nRIGA can be implemented in such way that it runs in polynomial time while\nasking no more than a polynomial number of queries. The method is tested on the\nmulti-objective knapsack and traveling salesman problems. For several\nperformance indicators (computation times, gap to optimality and number of\nqueries), RIGA obtains better results than state-of-the-art algorithms.\n", "date": "2023-11-10", "categories": []} {"id": "2303.16258", "title": "Optimisation via encodings: a renormalisation group perspective", "abstract": " Difficult, in particular NP-complete, optimization problems are traditionally\nsolved approximately using search heuristics. These are usually slowed down by\nthe rugged landscapes encountered, because local minima arrest the search\nprocess. Cover-encoding maps were devised to circumvent this problem by\ntransforming the original landscape to one that is free of local minima and\nenriched in near-optimal solutions. By definition, these involve the mapping of\nthe original (larger) search space into smaller subspaces, by processes that\ntypically amount to a form of coarse-graining. In this paper, we explore the\ndetails of this coarse-graining using formal arguments, as well as concrete\nexamples of cover-encoding maps, that are investigated analytically as well as\ncomputationally. Our results strongly suggest that the coarse-graining involved\nin cover-encoding maps bears a strong resemblance to that encountered in\nrenormalisation group schemes. Given the apparently disparate nature of these\ntwo formalisms, these strong similarities are rather startling, and suggest\ndeep mathematical underpinnings that await further exploration.\n", "date": "2023-03-28", "categories": []} {"id": "2310.05309", "title": "Optimizing Solution-Samplers for Combinatorial Problems: The Landscape\n of Policy-Gradient Methods", "abstract": " Deep Neural Networks and Reinforcement Learning methods have empirically\nshown great promise in tackling challenging combinatorial problems. In those\nmethods a deep neural network is used as a solution generator which is then\ntrained by gradient-based methods (e.g., policy gradient) to successively\nobtain better solution distributions. In this work we introduce a novel\ntheoretical framework for analyzing the effectiveness of such methods. We ask\nwhether there exist generative models that (i) are expressive enough to\ngenerate approximately optimal solutions; (ii) have a tractable, i.e,\npolynomial in the size of the input, number of parameters; (iii) their\noptimization landscape is benign in the sense that it does not contain\nsub-optimal stationary points. Our main contribution is a positive answer to\nthis question. Our result holds for a broad class of combinatorial problems\nincluding Max- and Min-Cut, Max-$k$-CSP, Maximum-Weight-Bipartite-Matching, and\nthe Traveling Salesman Problem. As a byproduct of our analysis we introduce a\nnovel regularization process over vanilla gradient descent and provide\ntheoretical and experimental evidence that it helps address vanishing-gradient\nissues and escape bad stationary points.\n", "date": "2023-10-08", "categories": []} {"id": "2311.03685", "title": "Dynamic Non-monotone Submodular Maximization", "abstract": " Maximizing submodular functions has been increasingly used in many\napplications of machine learning, such as data summarization, recommendation\nsystems, and feature selection. Moreover, there has been a growing interest in\nboth submodular maximization and dynamic algorithms. In 2020, Monemizadeh and\nLattanzi, Mitrovic, Norouzi{-}Fard, Tarnawski, and Zadimoghaddam initiated\ndeveloping dynamic algorithms for the monotone submodular maximization problem\nunder the cardinality constraint $k$. Recently, there have been some\nimprovements on the topic made by Banihashem, Biabani, Goudarzi, Hajiaghayi,\nJabbarzade, and Monemizadeh. In 2022, Chen and Peng studied the complexity of\nthis problem and raised an important open question: \"Can we extend [fully\ndynamic] results (algorithm or hardness) to non-monotone submodular\nmaximization?\". We affirmatively answer their question by demonstrating a\nreduction from maximizing a non-monotone submodular function under the\ncardinality constraint $k$ to maximizing a monotone submodular function under\nthe same constraint. Through this reduction, we obtain the first dynamic\nalgorithms to solve the non-monotone submodular maximization problem under the\ncardinality constraint $k$. Our algorithms maintain an\n$(8+\\epsilon)$-approximate of the solution and use expected amortized\n$O(\\epsilon^{-3}k^3\\log^3(n)\\log(k))$ or $O(\\epsilon^{-1}k^2\\log^3(k))$ oracle\nqueries per update, respectively. Furthermore, we showcase the benefits of our\ndynamic algorithm for video summarization and max-cut problems on several\nreal-world data sets.\n", "date": "2023-11-07", "categories": []} {"id": "2309.14032", "title": "DeepACO: Neural-enhanced Ant Systems for Combinatorial Optimization", "abstract": " Ant Colony Optimization (ACO) is a meta-heuristic algorithm that has been\nsuccessfully applied to various Combinatorial Optimization Problems (COPs).\nTraditionally, customizing ACO for a specific problem requires the expert\ndesign of knowledge-driven heuristics. In this paper, we propose DeepACO, a\ngeneric framework that leverages deep reinforcement learning to automate\nheuristic designs. DeepACO serves to strengthen the heuristic measures of\nexisting ACO algorithms and dispense with laborious manual design in future ACO\napplications. As a neural-enhanced meta-heuristic, DeepACO consistently\noutperforms its ACO counterparts on eight COPs using a single neural\narchitecture and a single set of hyperparameters. As a Neural Combinatorial\nOptimization method, DeepACO performs better than or on par with\nproblem-specific methods on canonical routing problems. Our code is publicly\navailable at https://github.com/henry-yeh/DeepACO.\n", "date": "2023-09-25", "categories": []} {"id": "2311.02101", "title": "Solving MaxSAT with Matrix Multiplication", "abstract": " We propose an incomplete algorithm for Maximum Satisfiability (MaxSAT)\nspecifically designed to run on neural network accelerators such as GPUs and\nTPUs. Given a MaxSAT problem instance in conjunctive normal form, our procedure\nconstructs a Restricted Boltzmann Machine (RBM) with an equilibrium\ndistribution wherein the probability of a Boolean assignment is exponential in\nthe number of clauses it satisfies. Block Gibbs sampling is used to\nstochastically search the space of assignments with parallel Markov chains.\nSince matrix multiplication is the main computational primitive for block Gibbs\nsampling in an RBM, our approach leads to an elegantly simple algorithm (40\nlines of JAX) well-suited for neural network accelerators. Theoretical results\nabout RBMs guarantee that the required number of visible and hidden units of\nthe RBM scale only linearly with the number of variables and constant-sized\nclauses in the MaxSAT instance, ensuring that the computational cost of a Gibbs\nstep scales reasonably with the instance size. Search throughput can be\nincreased by batching parallel chains within a single accelerator as well as by\ndistributing them across multiple accelerators. As a further enhancement, a\nheuristic based on unit propagation running on CPU is periodically applied to\nthe sampled assignments. Our approach, which we term RbmSAT, is a new design\npoint in the algorithm-hardware co-design space for MaxSAT. We present timed\nresults on a subset of problem instances from the annual MaxSAT Evaluation's\nIncomplete Unweighted Track for the years 2018 to 2021. When allotted the same\nrunning time and CPU compute budget (but no TPUs), RbmSAT outperforms other\nparticipating solvers on problems drawn from three out of the four years'\ncompetitions. Given the same running time on a TPU cluster for which RbmSAT is\nuniquely designed, it outperforms all solvers on problems drawn from all four\nyears.\n", "date": "2023-11-01", "categories": []} {"id": "2307.08964", "title": "Landscape Surrogate: Learning Decision Losses for Mathematical\n Optimization Under Partial Information", "abstract": " Recent works in learning-integrated optimization have shown promise in\nsettings where the optimization problem is only partially observed or where\ngeneral-purpose optimizers perform poorly without expert tuning. By learning an\noptimizer $\\mathbf{g}$ to tackle these challenging problems with $f$ as the\nobjective, the optimization process can be substantially accelerated by\nleveraging past experience. The optimizer can be trained with supervision from\nknown optimal solutions or implicitly by optimizing the compound function\n$f\\circ \\mathbf{g}$. The implicit approach may not require optimal solutions as\nlabels and is capable of handling problem uncertainty; however, it is slow to\ntrain and deploy due to frequent calls to optimizer $\\mathbf{g}$ during both\ntraining and testing. The training is further challenged by sparse gradients of\n$\\mathbf{g}$, especially for combinatorial solvers. To address these\nchallenges, we propose using a smooth and learnable Landscape Surrogate $M$ as\na replacement for $f\\circ \\mathbf{g}$. This surrogate, learnable by neural\nnetworks, can be computed faster than the solver $\\mathbf{g}$, provides dense\nand smooth gradients during training, can generalize to unseen optimization\nproblems, and is efficiently learned via alternating optimization. We test our\napproach on both synthetic problems, including shortest path and\nmultidimensional knapsack, and real-world problems such as portfolio\noptimization, achieving comparable or superior objective values compared to\nstate-of-the-art baselines while reducing the number of calls to $\\mathbf{g}$.\nNotably, our approach outperforms existing methods for computationally\nexpensive high-dimensional problems.\n", "date": "2023-07-18", "categories": []} {"id": "2309.05183", "title": "Data Summarization beyond Monotonicity: Non-monotone Two-Stage\n Submodular Maximization", "abstract": " The objective of a two-stage submodular maximization problem is to reduce the\nground set using provided training functions that are submodular, with the aim\nof ensuring that optimizing new objective functions over the reduced ground set\nyields results comparable to those obtained over the original ground set. This\nproblem has applications in various domains including data summarization.\nExisting studies often assume the monotonicity of the objective function,\nwhereas our work pioneers the extension of this research to accommodate\nnon-monotone submodular functions. We have introduced the first constant-factor\napproximation algorithms for this more general case.\n", "date": "2023-09-11", "categories": []} {"id": "2311.00971", "title": "An Integrated Framework Integrating Monte Carlo Tree Search and\n Supervised Learning for Train Timetabling Problem", "abstract": " The single-track railway train timetabling problem (TTP) is an important and\ncomplex problem. This article proposes an integrated Monte Carlo Tree Search\n(MCTS) computing framework that combines heuristic methods, unsupervised\nlearning methods, and supervised learning methods for solving TTP in discrete\naction spaces. This article first describes the mathematical model and\nsimulation system dynamics of TTP, analyzes the characteristics of the solution\nfrom the perspective of MCTS, and proposes some heuristic methods to improve\nMCTS. This article considers these methods as planners in the proposed\nframework. Secondly, this article utilizes deep convolutional neural networks\nto approximate the value of nodes and further applies them to the MCTS search\nprocess, referred to as learners. The experiment shows that the proposed\nheuristic MCTS method is beneficial for solving TTP; The algorithm framework\nthat integrates planners and learners can improve the data efficiency of\nsolving TTP; The proposed method provides a new paradigm for solving TTP.\n", "date": "2023-11-02", "categories": []} {"id": "2304.11839", "title": "Local Energy Distribution Based Hyperparameter Determination for\n Stochastic Simulated Annealing", "abstract": " This paper presents a local energy distribution based hyperparameter\ndetermination for stochastic simulated annealing (SSA). SSA is capable of\nsolving combinatorial optimization problems faster than typical simulated\nannealing (SA), but requires a time-consuming hyperparameter search. The\nproposed method determines hyperparameters based on the local energy\ndistributions of spins (probabilistic bits). The spin is a basic computing\nelement of SSA and is graphically connected to other spins with its weights.\nThe distribution of the local energy can be estimated based on the central\nlimit theorem (CLT). The CLT-based normal distribution is used to determine the\nhyperparameters, which reduces the time complexity for hyperparameter search\nfrom O(n^3) of the conventional method to O(1). The performance of SSA with the\ndetermined hyperparameters is evaluated on the Gset and K2000 benchmarks for\nmaximum-cut problems. The results show that the proposed method achieves mean\ncut values of approximately 98% of the best-known cut values.\n", "date": "2023-04-24", "categories": []} {"id": "2302.09166", "title": "Machine Learning for Cutting Planes in Integer Programming: A Survey", "abstract": " We survey recent work on machine learning (ML) techniques for selecting\ncutting planes (or cuts) in mixed-integer linear programming (MILP). Despite\nthe availability of various classes of cuts, the task of choosing a set of cuts\nto add to the linear programming (LP) relaxation at a given node of the\nbranch-and-bound (B&B) tree has defied both formal and heuristic solutions to\ndate. ML offers a promising approach for improving the cut selection process by\nusing data to identify promising cuts that accelerate the solution of MILP\ninstances. This paper presents an overview of the topic, highlighting recent\nadvances in the literature, common approaches to data collection, evaluation,\nand ML model architectures. We analyze the empirical results in the literature\nin an attempt to quantify the progress that has been made and conclude by\nsuggesting avenues for future research.\n", "date": "2023-02-17", "categories": []} {"id": "2310.07156", "title": "Solving Travelling Thief Problems using Coordination Based Methods", "abstract": " A travelling thief problem (TTP) is a proxy to real-life problems such as\npostal collection. TTP comprises an entanglement of a travelling salesman\nproblem (TSP) and a knapsack problem (KP) since items of KP are scattered over\ncities of TSP, and a thief has to visit cities to collect items. In TTP, city\nselection and item selection decisions need close coordination since the\nthief's travelling speed depends on the knapsack's weight and the order of\nvisiting cities affects the order of item collection. Existing TTP solvers deal\nwith city selection and item selection separately, keeping decisions for one\ntype unchanged while dealing with the other type. This separation essentially\nmeans very poor coordination between two types of decision. In this paper, we\nfirst show that a simple local search based coordination approach does not work\nin TTP. Then, to address the aforementioned problems, we propose a human\ndesigned coordination heuristic that makes changes to collection plans during\nexploration of cyclic tours. We further propose another human designed\ncoordination heuristic that explicitly exploits the cyclic tours in item\nselections during collection plan exploration. Lastly, we propose a machine\nlearning based coordination heuristic that captures characteristics of the two\nhuman designed coordination heuristics. Our proposed coordination based\napproaches help our TTP solver significantly outperform existing\nstate-of-the-art TTP solvers on a set of benchmark problems. Our solver is\nnamed Cooperation Coordination (CoCo) and its source code is available from\nhttps://github.com/majid75/CoCo\n", "date": "2023-10-11", "categories": []} {"id": "2310.19814", "title": "Learning Gradient Fields for Scalable and Generalizable Irregular\n Packing", "abstract": " The packing problem, also known as cutting or nesting, has diverse\napplications in logistics, manufacturing, layout design, and atlas generation.\nIt involves arranging irregularly shaped pieces to minimize waste while\navoiding overlap. Recent advances in machine learning, particularly\nreinforcement learning, have shown promise in addressing the packing problem.\nIn this work, we delve deeper into a novel machine learning-based approach that\nformulates the packing problem as conditional generative modeling. To tackle\nthe challenges of irregular packing, including object validity constraints and\ncollision avoidance, our method employs the score-based diffusion model to\nlearn a series of gradient fields. These gradient fields encode the\ncorrelations between constraint satisfaction and the spatial relationships of\npolygons, learned from teacher examples. During the testing phase, packing\nsolutions are generated using a coarse-to-fine refinement mechanism guided by\nthe learned gradient fields. To enhance packing feasibility and optimality, we\nintroduce two key architectural designs: multi-scale feature extraction and\ncoarse-to-fine relation extraction. We conduct experiments on two typical\nindustrial packing domains, considering translations only. Empirically, our\napproach demonstrates spatial utilization rates comparable to, or even\nsurpassing, those achieved by the teacher algorithm responsible for training\ndata generation. Additionally, it exhibits some level of generalization to\nshape variations. We are hopeful that this method could pave the way for new\npossibilities in solving the packing problem.\n", "date": "2023-10-18", "categories": []} {"id": "2310.19920", "title": "Solving a Class of Cut-Generating Linear Programs via Machine Learning", "abstract": " Cut-generating linear programs (CGLPs) play a key role as a separation oracle\nto produce valid inequalities for the feasible region of mixed-integer\nprograms. When incorporated inside branch-and-bound, the cutting planes\nobtained from CGLPs help to tighten relaxations and improve dual bounds.\nHowever, running the CGLPs at the nodes of the branch-and-bound tree is\ncomputationally cumbersome due to the large number of node candidates and the\nlack of a priori knowledge on which nodes admit useful cutting planes. As a\nresult, CGLPs are often avoided at default settings of branch-and-cut\nalgorithms despite their potential impact on improving dual bounds. In this\npaper, we propose a novel framework based on machine learning to approximate\nthe optimal value of a CGLP class that determines whether a cutting plane can\nbe generated at a node of the branch-and-bound tree. Translating the CGLP as an\nindicator function of the objective function vector, we show that it can be\napproximated through conventional data classification techniques. We provide a\nsystematic procedure to efficiently generate training data sets for the\ncorresponding classification problem based on the CGLP structure. We conduct\ncomputational experiments on benchmark instances using classification methods\nsuch as logistic regression. These results suggest that the approximate CGLP\nobtained from classification can improve the solution time compared to that of\nconventional cutting plane methods. Our proposed framework can be efficiently\napplied to a large number of nodes in the branch-and-bound tree to identify the\nbest candidates for adding a cut.\n", "date": "2023-10-30", "categories": []} {"id": "2310.19990", "title": "Unveiling the Limits of Learned Local Search Heuristics: Are You the\n Mightiest of the Meek?", "abstract": " In recent years, combining neural networks with local search heuristics has\nbecome popular in the field of combinatorial optimization. Despite its\nconsiderable computational demands, this approach has exhibited promising\noutcomes with minimal manual engineering. However, we have identified three\ncritical limitations in the empirical evaluation of these integration attempts.\nFirstly, instances with moderate complexity and weak baselines pose a challenge\nin accurately evaluating the effectiveness of learning-based approaches.\nSecondly, the absence of an ablation study makes it difficult to quantify and\nattribute improvements accurately to the deep learning architecture. Lastly,\nthe generalization of learned heuristics across diverse distributions remains\nunderexplored. In this study, we conduct a comprehensive investigation into\nthese identified limitations. Surprisingly, we demonstrate that a simple\nlearned heuristic based on Tabu Search surpasses state-of-the-art (SOTA)\nlearned heuristics in terms of performance and generalizability. Our findings\nchallenge prevailing assumptions and open up exciting avenues for future\nresearch and innovation in combinatorial optimization.\n", "date": "2023-10-30", "categories": []} {"id": "2306.17283", "title": "A Neural Separation Algorithm for the Rounded Capacity Inequalities", "abstract": " The cutting plane method is a key technique for successful branch-and-cut and\nbranch-price-and-cut algorithms that find the exact optimal solutions for\nvarious vehicle routing problems (VRPs). Among various cuts, the rounded\ncapacity inequalities (RCIs) are the most fundamental. To generate RCIs, we\nneed to solve the separation problem, whose exact solution takes a long time to\nobtain; therefore, heuristic methods are widely used. We design a\nlearning-based separation heuristic algorithm with graph coarsening that learns\nthe solutions of the exact separation problem with a graph neural network\n(GNN), which is trained with small instances of 50 to 100 customers. We embed\nour separation algorithm within the cutting plane method to find a lower bound\nfor the capacitated VRP (CVRP) with up to 1,000 customers. We compare the\nperformance of our approach with CVRPSEP, a popular separation software package\nfor various cuts used in solving VRPs. Our computational results show that our\napproach finds better lower bounds than CVRPSEP for large-scale problems with\n400 or more customers, while CVRPSEP shows strong competency for problems with\nless than 400 customers.\n", "date": "2023-06-29", "categories": []} {"id": "2310.18370", "title": "New Boolean satisfiability problem heuristic strategy: Minimal Positive\n Negative Product Strategy", "abstract": " This study presents a novel heuristic algorithm called the \"Minimal Positive\nNegative Product Strategy\" to guide the CDCL algorithm in solving the Boolean\nsatisfiability problem. It provides a mathematical explanation for the\nsuperiority of this algorithm over widely used heuristics such as the Dynamic\nLargest Individual Sum (DLIS) and the Variable State Independent Decaying Sum\n(VSIDS). Experimental results further confirm the effectiveness of this\nheuristic strategy in problem-solving.\n", "date": "2023-10-26", "categories": []} {"id": "2207.08189", "title": "Supplementing Recurrent Neural Networks with Annealing to Solve\n Combinatorial Optimization Problems", "abstract": " Combinatorial optimization problems can be solved by heuristic algorithms\nsuch as simulated annealing (SA) which aims to find the optimal solution within\na large search space through thermal fluctuations. The algorithm generates new\nsolutions through Markov-chain Monte Carlo techniques. This sampling scheme can\nresult in severe limitations, such as slow convergence and a tendency to stay\nwithin the same local search space at small temperatures. To overcome these\nshortcomings, we use the variational classical annealing (VCA) framework that\ncombines autoregressive recurrent neural networks (RNNs) with traditional\nannealing to sample solutions that are uncorrelated. In this paper, we\ndemonstrate the potential of using VCA as an approach to solving real-world\noptimization problems. We explore VCA's performance in comparison with SA at\nsolving three popular optimization problems: the maximum cut problem (Max-Cut),\nthe nurse scheduling problem (NSP), and the traveling salesman problem (TSP).\nFor all three problems, we find that VCA outperforms SA on average in the\nasymptotic limit by one or more orders of magnitude in terms of relative error.\nInterestingly, we reach large system sizes of up to $256$ cities for the TSP.\nWe also conclude that in the best case scenario, VCA can serve as a great\nalternative when SA fails to find the optimal solution.\n", "date": "2022-07-17", "categories": []} {"id": "2310.18264", "title": "Learning to Search Feasible and Infeasible Regions of Routing Problems\n with Flexible Neural k-Opt", "abstract": " In this paper, we present Neural k-Opt (NeuOpt), a novel learning-to-search\n(L2S) solver for routing problems. It learns to perform flexible k-opt\nexchanges based on a tailored action factorization method and a customized\nrecurrent dual-stream decoder. As a pioneering work to circumvent the pure\nfeasibility masking scheme and enable the autonomous exploration of both\nfeasible and infeasible regions, we then propose the Guided Infeasible Region\nExploration (GIRE) scheme, which supplements the NeuOpt policy network with\nfeasibility-related features and leverages reward shaping to steer\nreinforcement learning more effectively. Additionally, we equip NeuOpt with\nDynamic Data Augmentation (D2A) for more diverse searches during inference.\nExtensive experiments on the Traveling Salesman Problem (TSP) and Capacitated\nVehicle Routing Problem (CVRP) demonstrate that our NeuOpt not only\nsignificantly outstrips existing (masking-based) L2S solvers, but also\nshowcases superiority over the learning-to-construct (L2C) and\nlearning-to-predict (L2P) solvers. Notably, we offer fresh perspectives on how\nneural solvers can handle VRP constraints. Our code is available:\nhttps://github.com/yining043/NeuOpt.\n", "date": "2023-10-27", "categories": []} {"id": "2310.15195", "title": "Neural Multi-Objective Combinatorial Optimization with Diversity\n Enhancement", "abstract": " Most of existing neural methods for multi-objective combinatorial\noptimization (MOCO) problems solely rely on decomposition, which often leads to\nrepetitive solutions for the respective subproblems, thus a limited Pareto set.\nBeyond decomposition, we propose a novel neural heuristic with diversity\nenhancement (NHDE) to produce more Pareto solutions from two perspectives. On\nthe one hand, to hinder duplicated solutions for different subproblems, we\npropose an indicator-enhanced deep reinforcement learning method to guide the\nmodel, and design a heterogeneous graph attention mechanism to capture the\nrelations between the instance graph and the Pareto front graph. On the other\nhand, to excavate more solutions in the neighborhood of each subproblem, we\npresent a multiple Pareto optima strategy to sample and preserve desirable\nsolutions. Experimental results on classic MOCO problems show that our NHDE is\nable to generate a Pareto front with higher diversity, thereby achieving\nsuperior overall performance. Moreover, our NHDE is generic and can be applied\nto different neural methods for MOCO.\n", "date": "2023-10-22", "categories": []} {"id": "2310.15196", "title": "Efficient Meta Neural Heuristic for Multi-Objective Combinatorial\n Optimization", "abstract": " Recently, neural heuristics based on deep reinforcement learning have\nexhibited promise in solving multi-objective combinatorial optimization\nproblems (MOCOPs). However, they are still struggling to achieve high learning\nefficiency and solution quality. To tackle this issue, we propose an efficient\nmeta neural heuristic (EMNH), in which a meta-model is first trained and then\nfine-tuned with a few steps to solve corresponding single-objective\nsubproblems. Specifically, for the training process, a (partial)\narchitecture-shared multi-task model is leveraged to achieve parallel learning\nfor the meta-model, so as to speed up the training; meanwhile, a scaled\nsymmetric sampling method with respect to the weight vectors is designed to\nstabilize the training. For the fine-tuning process, an efficient hierarchical\nmethod is proposed to systematically tackle all the subproblems. Experimental\nresults on the multi-objective traveling salesman problem (MOTSP),\nmulti-objective capacitated vehicle routing problem (MOCVRP), and\nmulti-objective knapsack problem (MOKP) show that, EMNH is able to outperform\nthe state-of-the-art neural heuristics in terms of solution quality and\nlearning efficiency, and yield competitive solutions to the strong traditional\nheuristics while consuming much shorter time.\n", "date": "2023-10-22", "categories": []} {"id": "2307.00663", "title": "Solving Multi-Agent Target Assignment and Path Finding with a Single\n Constraint Tree", "abstract": " Combined Target-Assignment and Path-Finding problem (TAPF) requires\nsimultaneously assigning targets to agents and planning collision-free paths\nfor agents from their start locations to their assigned targets. As a leading\napproach to address TAPF, Conflict-Based Search with Target Assignment (CBS-TA)\nleverages both K-best target assignments to create multiple search trees and\nConflict-Based Search (CBS) to resolve collisions in each search tree. While\nbeing able to find an optimal solution, CBS-TA suffers from scalability due to\nthe duplicated collision resolution in multiple trees and the expensive\ncomputation of K-best assignments. We therefore develop Incremental Target\nAssignment CBS (ITA-CBS) to bypass these two computational bottlenecks. ITA-CBS\ngenerates only a single search tree and avoids computing K-best assignments by\nincrementally computing new 1-best assignments during the search. We show that,\nin theory, ITA-CBS is guaranteed to find an optimal solution and, in practice,\nis computationally efficient.\n", "date": "2023-07-02", "categories": []} {"id": "2310.07980", "title": "GRASP: Accelerating Shortest Path Attacks via Graph Attention", "abstract": " Recent advances in machine learning (ML) have shown promise in aiding and\naccelerating classical combinatorial optimization algorithms. ML-based speed\nups that aim to learn in an end to end manner (i.e., directly output the\nsolution) tend to trade off run time with solution quality. Therefore,\nsolutions that are able to accelerate existing solvers while maintaining their\nperformance guarantees, are of great interest. We consider an APX-hard problem,\nwhere an adversary aims to attack shortest paths in a graph by removing the\nminimum number of edges. We propose the GRASP algorithm: Graph Attention\nAccelerated Shortest Path Attack, an ML aided optimization algorithm that\nachieves run times up to 10x faster, while maintaining the quality of solution\ngenerated. GRASP uses a graph attention network to identify a smaller subgraph\ncontaining the combinatorial solution, thus effectively reducing the input\nproblem size. Additionally, we demonstrate how careful representation of the\ninput graph, including node features that correlate well with the optimization\ntask, can highlight important structure in the optimization solution.\n", "date": "2023-10-12", "categories": []} {"id": "2310.13261", "title": "DIG-MILP: a Deep Instance Generator for Mixed-Integer Linear Programming\n with Feasibility Guarantee", "abstract": " Mixed-integer linear programming (MILP) stands as a notable NP-hard problem\npivotal to numerous crucial industrial applications. The development of\neffective algorithms, the tuning of solvers, and the training of machine\nlearning models for MILP resolution all hinge on access to extensive, diverse,\nand representative data. Yet compared to the abundant naturally occurring data\nin image and text realms, MILP is markedly data deficient, underscoring the\nvital role of synthetic MILP generation. We present DIG-MILP, a deep generative\nframework based on variational auto-encoder (VAE), adept at extracting\ndeep-level structural features from highly limited MILP data and producing\ninstances that closely mirror the target data. Notably, by leveraging the MILP\nduality, DIG-MILP guarantees a correct and complete generation space as well as\nensures the boundedness and feasibility of the generated instances. Our\nempirical study highlights the novelty and quality of the instances generated\nby DIG-MILP through two distinct downstream tasks: (S1) Data sharing, where\nsolver solution times correlate highly positive between original and\nDIG-MILP-generated instances, allowing data sharing for solver tuning without\npublishing the original data; (S2) Data Augmentation, wherein the\nDIG-MILP-generated instances bolster the generalization performance of machine\nlearning models tasked with resolving MILP problems.\n", "date": "2023-10-20", "categories": []} {"id": "2310.13311", "title": "Non-Negative Spherical Relaxations for Universe-Free Multi-Matching and\n Clustering", "abstract": " We propose a novel non-negative spherical relaxation for optimization\nproblems over binary matrices with injectivity constraints, which in particular\nhas applications in multi-matching and clustering. We relax respective binary\nmatrix constraints to the (high-dimensional) non-negative sphere. To optimize\nour relaxed problem, we use a conditional power iteration method to iteratively\nimprove the objective function, while at same time sweeping over a continuous\nscalar parameter that is (indirectly) related to the universe size (or number\nof clusters). Opposed to existing procedures that require to fix the integer\nuniverse size before optimization, our method automatically adjusts the\nanalogous continuous parameter. Furthermore, while our approach shares\nsimilarities with spectral multi-matching and spectral clustering, our\nformulation has the strong advantage that we do not rely on additional\npost-processing procedures to obtain binary results. Our method shows\ncompelling results in various multi-matching and clustering settings, even when\ncompared to methods that use the ground truth universe size (or number of\nclusters).\n", "date": "2023-10-20", "categories": []} {"id": "2310.09986", "title": "On Statistical Learning of Branch and Bound for Vehicle Routing\n Optimization", "abstract": " Recently, machine learning of the branch and bound algorithm has shown\npromise in approximating competent solutions to NP-hard problems. In this\npaper, we utilize and comprehensively compare the outcomes of three neural\nnetworks--graph convolutional neural network (GCNN), GraphSAGE, and graph\nattention network (GAT)--to solve the capacitated vehicle routing problem. We\ntrain these neural networks to emulate the decision-making process of the\ncomputationally expensive Strong Branching strategy. The neural networks are\ntrained on six instances with distinct topologies from the CVRPLIB and\nevaluated on eight additional instances. Moreover, we reduced the minimum\nnumber of vehicles required to solve a CVRP instance to a bin-packing problem,\nwhich was addressed in a similar manner. Through rigorous experimentation, we\nfound that this approach can match or improve upon the performance of the\nbranch and bound algorithm with the Strong Branching strategy while requiring\nsignificantly less computational time. The source code that corresponds to our\nresearch findings and methodology is readily accessible and available for\nreference at the following web address: https://isotlaboratory.github.io/ml4vrp\n", "date": "2023-10-15", "categories": []} {"id": "2310.10667", "title": "Enhancing Network Resilience through Machine Learning-powered Graph\n Combinatorial Optimization: Applications in Cyber Defense and Information\n Diffusion", "abstract": " With the burgeoning advancements of computing and network communication\ntechnologies, network infrastructures and their application environments have\nbecome increasingly complex. Due to the increased complexity, networks are more\nprone to hardware faults and highly susceptible to cyber-attacks. Therefore,\nfor rapidly growing network-centric applications, network resilience is\nessential to minimize the impact of attacks and to ensure that the network\nprovides an acceptable level of services during attacks, faults or disruptions.\nIn this regard, this thesis focuses on developing effective approaches for\nenhancing network resilience. Existing approaches for enhancing network\nresilience emphasize on determining bottleneck nodes and edges in the network\nand designing proactive responses to safeguard the network against attacks.\nHowever, existing solutions generally consider broader application domains and\npossess limited applicability when applied to specific application areas such\nas cyber defense and information diffusion, which are highly popular\napplication domains among cyber attackers.\n This thesis aims to design effective, efficient and scalable techniques for\ndiscovering bottleneck nodes and edges in the network to enhance network\nresilience in cyber defense and information diffusion application domains. We\nfirst investigate a cyber defense graph optimization problem, i.e., hardening\nactive directory systems by discovering bottleneck edges in the network. We\nthen study the problem of identifying bottleneck structural hole spanner nodes,\nwhich are crucial for information diffusion in the network. We transform both\nproblems into graph-combinatorial optimization problems and design machine\nlearning based approaches for discovering bottleneck points vital for enhancing\nnetwork resilience.\n", "date": "2023-09-22", "categories": []} {"id": "2310.09686", "title": "Enhancing Column Generation by Reinforcement Learning-Based\n Hyper-Heuristic for Vehicle Routing and Scheduling Problems", "abstract": " Column generation (CG) is a vital method to solve large-scale problems by\ndynamically generating variables. It has extensive applications in common\ncombinatorial optimization, such as vehicle routing and scheduling problems,\nwhere each iteration step requires solving an NP-hard constrained shortest path\nproblem. Although some heuristic methods for acceleration already exist, they\nare not versatile enough to solve different problems. In this work, we propose\na reinforcement learning-based hyper-heuristic framework, dubbed RLHH, to\nenhance the performance of CG. RLHH is a selection module embedded in CG to\naccelerate convergence and get better integer solutions. In each CG iteration,\nthe RL agent selects a low-level heuristic to construct a reduced network only\ncontaining the edges with a greater chance of being part of the optimal\nsolution. In addition, we specify RLHH to solve two typical combinatorial\noptimization problems: Vehicle Routing Problem with Time Windows (VRPTW) and\nBus Driver Scheduling Problem (BDSP). The total cost can be reduced by up to\n27.9\\% in VRPTW and 15.4\\% in BDSP compared to the best lower-level heuristic\nin our tested scenarios, within equivalent or even less computational time. The\nproposed RLHH is the first RL-based CG method that outperforms traditional\napproaches in terms of solution quality, which can promote the application of\nCG in combinatorial optimization.\n", "date": "2023-10-15", "categories": []} {"id": "2310.10603", "title": "Exploring the Power of Graph Neural Networks in Solving Linear\n Optimization Problems", "abstract": " Recently, machine learning, particularly message-passing graph neural\nnetworks (MPNNs), has gained traction in enhancing exact optimization\nalgorithms. For example, MPNNs speed up solving mixed-integer optimization\nproblems by imitating computational intensive heuristics like strong branching,\nwhich entails solving multiple linear optimization problems (LPs). Despite the\nempirical success, the reasons behind MPNNs' effectiveness in emulating linear\noptimization remain largely unclear. Here, we show that MPNNs can simulate\nstandard interior-point methods for LPs, explaining their practical success.\nFurthermore, we highlight how MPNNs can serve as a lightweight proxy for\nsolving LPs, adapting to a given problem instance distribution. Empirically, we\nshow that MPNNs solve LP relaxations of standard combinatorial optimization\nproblems close to optimality, often surpassing conventional solvers and\ncompeting approaches in solving time.\n", "date": "2023-10-16", "categories": []} {"id": "2310.09071", "title": "Online Relocating and Matching of Ride-Hailing Services: A Model-Based\n Modular Approach", "abstract": " This study proposes an innovative model-based modular approach (MMA) to\ndynamically optimize order matching and vehicle relocation in a ride-hailing\nplatform. MMA utilizes a two-layer and modular modeling structure. The upper\nlayer determines the spatial transfer patterns of vehicle flow within the\nsystem to maximize the total revenue of the current and future stages. With the\nguidance provided by the upper layer, the lower layer performs rapid\nvehicle-to-order matching and vehicle relocation. MMA is interpretable, and\nequipped with the customized and polynomial-time algorithm, which, as an online\norder-matching and vehicle-relocation algorithm, can scale past thousands of\nvehicles. We theoretically prove that the proposed algorithm can achieve the\nglobal optimum in stylized networks, while the numerical experiments based on\nboth the toy network and realistic dataset demonstrate that MMA is capable of\nachieving superior systematic performance compared to batch matching and\nreinforcement-learning based methods. Moreover, its modular and lightweight\nmodeling structure further enables it to achieve a high level of robustness\nagainst demand variation while maintaining a relatively low computational cost.\n", "date": "2023-10-13", "categories": []} {"id": "2206.06924", "title": "The Maximum Linear Arrangement Problem for trees under projectivity and\n planarity", "abstract": " A linear arrangement is a mapping $\\pi$ from the $n$ vertices of a graph $G$\nto $n$ distinct consecutive integers. Linear arrangements can be represented by\ndrawing the vertices along a horizontal line and drawing the edges as\nsemicircles above said line. In this setting, the length of an edge is defined\nas the absolute value of the difference between the positions of its two\nvertices in the arrangement, and the cost of an arrangement as the sum of all\nedge lengths. Here we study two variants of the Maximum Linear Arrangement\nproblem (MaxLA), which consists of finding an arrangement that maximizes the\ncost. In the planar variant for free trees, vertices have to be arranged in\nsuch a way that there are no edge crossings. In the projective variant for\nrooted trees, arrangements have to be planar and the root of the tree cannot be\ncovered by any edge. In this paper we present algorithms that are linear in\ntime and space to solve planar and projective MaxLA for trees. We also prove\nseveral properties of maximum projective and planar arrangements, and show that\ncaterpillar trees maximize planar MaxLA over all trees of a fixed size thereby\ngeneralizing a previous extremal result on trees.\n", "date": "2022-06-14", "categories": []} {"id": "2209.08901", "title": "Global Optimization for Cardinality-constrained Minimum Sum-of-Squares\n Clustering via Semidefinite Programming", "abstract": " The minimum sum-of-squares clustering (MSSC), or k-means type clustering, has\nbeen recently extended to exploit prior knowledge on the cardinality of each\ncluster. Such knowledge is used to increase performance as well as solution\nquality. In this paper, we propose a global optimization approach based on the\nbranch-and-cut technique to solve the cardinality-constrained MSSC. For the\nlower bound routine, we use the semidefinite programming (SDP) relaxation\nrecently proposed by Rujeerapaiboon et al. [SIAM J. Optim. 29(2), 1211-1239,\n(2019)]. However, this relaxation can be used in a branch-and-cut method only\nfor small-size instances. Therefore, we derive a new SDP relaxation that scales\nbetter with the instance size and the number of clusters. In both cases, we\nstrengthen the bound by adding polyhedral cuts. Benefiting from a tailored\nbranching strategy which enforces pairwise constraints, we reduce the\ncomplexity of the problems arising in the children nodes. For the upper bound,\ninstead, we present a local search procedure that exploits the solution of the\nSDP relaxation solved at each node. Computational results show that the\nproposed algorithm globally solves, for the first time, real-world instances of\nsize 10 times larger than those solved by state-of-the-art exact methods.\n", "date": "2022-09-19", "categories": []} {"id": "2310.08122", "title": "Core-sets for Fair and Diverse Data Summarization", "abstract": " We study core-set construction algorithms for the task of Diversity\nMaximization under fairness/partition constraint. Given a set of points $P$ in\na metric space partitioned into $m$ groups, and given $k_1,\\ldots,k_m$, the\ngoal of this problem is to pick $k_i$ points from each group $i$ such that the\noverall diversity of the $k=\\sum_i k_i$ picked points is maximized. We consider\ntwo natural diversity measures: sum-of-pairwise distances and\nsum-of-nearest-neighbor distances, and show improved core-set construction\nalgorithms with respect to these measures. More precisely, we show the first\nconstant factor core-set w.r.t. sum-of-pairwise distances whose size is\nindependent of the size of the dataset and the aspect ratio. Second, we show\nthe first core-set w.r.t. the sum-of-nearest-neighbor distances. Finally, we\nrun several experiments showing the effectiveness of our core-set approach. In\nparticular, we apply constrained diversity maximization to summarize a set of\ntimed messages that takes into account the messages' recency. Specifically, the\nsummary should include more recent messages compared to older ones. This is a\nreal task in one of the largest communication platforms, affecting the\nexperience of hundreds of millions daily active users. By utilizing our\ncore-set method for this task, we achieve a 100x speed-up while losing the\ndiversity by only a few percent. Moreover, our approach allows us to improve\nthe space usage of the algorithm in the streaming setting.\n", "date": "2023-10-12", "categories": []} {"id": "2206.00383", "title": "Neural Improvement Heuristics for Graph Combinatorial Optimization\n Problems", "abstract": " Recent advances in graph neural network architectures and increased\ncomputation power have revolutionized the field of combinatorial optimization\n(CO). Among the proposed models for CO problems, Neural Improvement (NI) models\nhave been particularly successful. However, existing NI approaches are limited\nin their applicability to problems where crucial information is encoded in the\nedges, as they only consider node features and node-wise positional encodings.\nTo overcome this limitation, we introduce a novel NI model capable of handling\ngraph-based problems where information is encoded in the nodes, edges, or both.\nThe presented model serves as a fundamental component for hill-climbing-based\nalgorithms that guide the selection of neighborhood operations for each\niteration. Conducted experiments demonstrate that the proposed model can\nrecommend neighborhood operations that outperform conventional versions for the\nPreference Ranking Problem with a performance in the 99th percentile. We also\nextend the proposal to two well-known problems: the Traveling Salesman Problem\nand the Graph Partitioning Problem, recommending operations in the 98th and\n97th percentile, respectively.\n", "date": "2022-06-01", "categories": []} {"id": "2307.12065", "title": "Spectral Normalized-Cut Graph Partitioning with Fairness Constraints", "abstract": " Normalized-cut graph partitioning aims to divide the set of nodes in a graph\ninto $k$ disjoint clusters to minimize the fraction of the total edges between\nany cluster and all other clusters. In this paper, we consider a fair variant\nof the partitioning problem wherein nodes are characterized by a categorical\nsensitive attribute (e.g., gender or race) indicating membership to different\ndemographic groups. Our goal is to ensure that each group is approximately\nproportionally represented in each cluster while minimizing the normalized cut\nvalue. To resolve this problem, we propose a two-phase spectral algorithm\ncalled FNM. In the first phase, we add an augmented Lagrangian term based on\nour fairness criteria to the objective function for obtaining a fairer spectral\nnode embedding. Then, in the second phase, we design a rounding scheme to\nproduce $k$ clusters from the fair embedding that effectively trades off\nfairness and partition quality. Through comprehensive experiments on nine\nbenchmark datasets, we demonstrate the superior performance of FNM compared\nwith three baseline methods.\n", "date": "2023-07-22", "categories": []} {"id": "2310.04140", "title": "Routing Arena: A Benchmark Suite for Neural Routing Solvers", "abstract": " Neural Combinatorial Optimization has been researched actively in the last\neight years. Even though many of the proposed Machine Learning based approaches\nare compared on the same datasets, the evaluation protocol exhibits essential\nflaws and the selection of baselines often neglects State-of-the-Art Operations\nResearch approaches. To improve on both of these shortcomings, we propose the\nRouting Arena, a benchmark suite for Routing Problems that provides a seamless\nintegration of consistent evaluation and the provision of baselines and\nbenchmarks prevalent in the Machine Learning- and Operations Research field.\nThe proposed evaluation protocol considers the two most important evaluation\ncases for different applications: First, the solution quality for an a priori\nfixed time budget and secondly the anytime performance of the respective\nmethods. By setting the solution trajectory in perspective to a Best Known\nSolution and a Base Solver's solutions trajectory, we furthermore propose the\nWeighted Relative Average Performance (WRAP), a novel evaluation metric that\nquantifies the often claimed runtime efficiency of Neural Routing Solvers. A\ncomprehensive first experimental evaluation demonstrates that the most recent\nOperations Research solvers generate state-of-the-art results in terms of\nsolution quality and runtime efficiency when it comes to the vehicle routing\nproblem. Nevertheless, some findings highlight the advantages of neural\napproaches and motivate a shift in how neural solvers should be conceptualized.\n", "date": "2023-10-06", "categories": []} {"id": "2309.15517", "title": "Residual Scheduling: A New Reinforcement Learning Approach to Solving\n Job Shop Scheduling Problem", "abstract": " Job-shop scheduling problem (JSP) is a mathematical optimization problem\nwidely used in industries like manufacturing, and flexible JSP (FJSP) is also a\ncommon variant. Since they are NP-hard, it is intractable to find the optimal\nsolution for all cases within reasonable times. Thus, it becomes important to\ndevelop efficient heuristics to solve JSP/FJSP. A kind of method of solving\nscheduling problems is construction heuristics, which constructs scheduling\nsolutions via heuristics. Recently, many methods for construction heuristics\nleverage deep reinforcement learning (DRL) with graph neural networks (GNN). In\nthis paper, we propose a new approach, named residual scheduling, to solving\nJSP/FJSP. In this new approach, we remove irrelevant machines and jobs such as\nthose finished, such that the states include the remaining (or relevant)\nmachines and jobs only. Our experiments show that our approach reaches\nstate-of-the-art (SOTA) among all known construction heuristics on most\nwell-known open JSP and FJSP benchmarks. In addition, we also observe that even\nthough our model is trained for scheduling problems of smaller sizes, our\nmethod still performs well for scheduling problems of large sizes.\nInterestingly in our experiments, our approach even reaches zero gap for 49\namong 50 JSP instances whose job numbers are more than 150 on 20 machines.\n", "date": "2023-09-27", "categories": []} {"id": "2310.01604", "title": "Solving the Quadratic Assignment Problem using Deep Reinforcement\n Learning", "abstract": " The Quadratic Assignment Problem (QAP) is an NP-hard problem which has proven\nparticularly challenging to solve: unlike other combinatorial problems like the\ntraveling salesman problem (TSP), which can be solved to optimality for\ninstances with hundreds or even thousands of locations using advanced integer\nprogramming techniques, no methods are known to exactly solve QAP instances of\nsize greater than 30. Solving the QAP is nevertheless important because of its\nmany critical applications, such as electronic wiring design and facility\nlayout selection. We propose a method to solve the original Koopmans-Beckman\nformulation of the QAP using deep reinforcement learning. Our approach relies\non a novel double pointer network, which alternates between selecting a\nlocation in which to place the next facility and a facility to place in the\nprevious location. We train our model using A2C on a large dataset of synthetic\ninstances, producing solutions with no instance-specific retraining necessary.\nOut of sample, our solutions are on average within 7.5% of a high-quality local\nsearch baseline, and even outperform it on 1.2% of instances.\n", "date": "2023-10-02", "categories": []} {"id": "2301.03313", "title": "BQ-NCO: Bisimulation Quotienting for Efficient Neural Combinatorial\n Optimization", "abstract": " Despite the success of neural-based combinatorial optimization methods for\nend-to-end heuristic learning, out-of-distribution generalization remains a\nchallenge. In this paper, we present a novel formulation of Combinatorial\nOptimization Problems (COPs) as Markov Decision Processes (MDPs) that\neffectively leverages common symmetries of COPs to improve out-of-distribution\nrobustness. Starting from a direct MDP formulation of a constructive method, we\nintroduce a generic way to reduce the state space, based on Bisimulation\nQuotienting (BQ) in MDPs. Then, for COPs with a recursive nature, we specialize\nthe bisimulation and show how the reduced state exploits the symmetries of\nthese problems and facilitates MDP solving. Our approach is principled and we\nprove that an optimal policy for the proposed BQ-MDP actually solves the\nassociated COPs. We illustrate our approach on five classical problems: the\nEuclidean and Asymmetric Traveling Salesman, Capacitated Vehicle Routing,\nOrienteering and Knapsack Problems. Furthermore, for each problem, we introduce\na simple attention-based policy network for the BQ-MDPs, which we train by\nimitation of (near) optimal solutions of small instances from a single\ndistribution. We obtain new state-of-the-art results for the five COPs on both\nsynthetic and realistic benchmarks. Notably, in contrast to most existing\nneural approaches, our learned policies show excellent generalization\nperformance to much larger instances than seen during training, without any\nadditional search procedure.\n", "date": "2023-01-09", "categories": []} {"id": "2302.02506", "title": "Generating Dispatching Rules for the Interrupting Swap-Allowed Blocking\n Job Shop Problem Using Graph Neural Network and Reinforcement Learning", "abstract": " The interrupting swap-allowed blocking job shop problem (ISBJSSP) is a\ncomplex scheduling problem that is able to model many manufacturing planning\nand logistics applications realistically by addressing both the lack of storage\ncapacity and unforeseen production interruptions. Subjected to random\ndisruptions due to machine malfunction or maintenance, industry production\nsettings often choose to adopt dispatching rules to enable adaptive, real-time\nre-scheduling, rather than traditional methods that require costly\nre-computation on the new configuration every time the problem condition\nchanges dynamically. To generate dispatching rules for the ISBJSSP problem, we\nintroduce a dynamic disjunctive graph formulation characterized by nodes and\nedges subjected to continuous deletions and additions. This formulation enables\nthe training of an adaptive scheduler utilizing graph neural networks and\nreinforcement learning. Furthermore, a simulator is developed to simulate\ninterruption, swapping, and blocking in the ISBJSSP setting. Employing a set of\nreported benchmark instances, we conduct a detailed experimental study on\nISBJSSP instances with a range of machine shutdown probabilities to show that\nthe scheduling policies generated can outperform or are at least as competitive\nas existing dispatching rules with predetermined priority. This study shows\nthat the ISBJSSP, which requires real-time adaptive solutions, can be scheduled\nefficiently with the proposed method when production interruptions occur with\nrandom machine shutdowns.\n", "date": "2023-02-05", "categories": []} {"id": "2309.17089", "title": "Too Big, so Fail? -- Enabling Neural Construction Methods to Solve\n Large-Scale Routing Problems", "abstract": " In recent years new deep learning approaches to solve combinatorial\noptimization problems, in particular NP-hard Vehicle Routing Problems (VRP),\nhave been proposed. The most impactful of these methods are sequential neural\nconstruction approaches which are usually trained via reinforcement learning.\nDue to the high training costs of these models, they usually are trained on\nlimited instance sizes (e.g. serving 100 customers) and later applied to vastly\nlarger instance size (e.g. 2000 customers). By means of a systematic scale-up\nstudy we show that even state-of-the-art neural construction methods are\noutperformed by simple heuristics, failing to generalize to larger problem\ninstances. We propose to use the ruin recreate principle that alternates\nbetween completely destroying a localized part of the solution and then\nrecreating an improved variant. In this way, neural construction methods like\nPOMO are never applied to the global problem but just in the reconstruction\nstep, which only involves partial problems much closer in size to their\noriginal training instances. In thorough experiments on four datasets of\nvarying distributions and modalities we show that our neural ruin recreate\napproach outperforms alternative forms of improving construction methods such\nas sampling and beam search and in several experiments also advanced local\nsearch approaches.\n", "date": "2023-09-29", "categories": []} {"id": "2308.13978", "title": "A Graph Neural Network-Based QUBO-Formulated Hamiltonian-Inspired Loss\n Function for Combinatorial Optimization using Reinforcement Learning", "abstract": " Quadratic Unconstrained Binary Optimization (QUBO) is a generic technique to\nmodel various NP-hard combinatorial optimization problems in the form of binary\nvariables. The Hamiltonian function is often used to formulate QUBO problems\nwhere it is used as the objective function in the context of optimization.\nRecently, PI-GNN, a generic scalable framework, has been proposed to address\nthe Combinatorial Optimization (CO) problems over graphs based on a simple\nGraph Neural Network (GNN) architecture. Their novel contribution was a generic\nQUBO-formulated Hamiltonian-inspired loss function that was optimized using\nGNN. In this study, we address a crucial issue related to the aforementioned\nsetup especially observed in denser graphs. The reinforcement learning-based\nparadigm has also been widely used to address numerous CO problems. Here we\nalso formulate and empirically evaluate the compatibility of the\nQUBO-formulated Hamiltonian as the generic reward function in the Reinforcement\nLearning paradigm to directly integrate the actual node projection status\nduring training as the form of rewards. In our experiments, we observed up to\n44% improvement in the RL-based setup compared to the PI-GNN algorithm. Our\nimplementation can be found in\nhttps://github.com/rizveeredwan/learning-graph-structure.\n", "date": "2023-08-27", "categories": []} {"id": "2309.16413", "title": "Genetic Engineering Algorithm (GEA): An Efficient Metaheuristic\n Algorithm for Solving Combinatorial Optimization Problems", "abstract": " Genetic Algorithms (GAs) are known for their efficiency in solving\ncombinatorial optimization problems, thanks to their ability to explore diverse\nsolution spaces, handle various representations, exploit parallelism, preserve\ngood solutions, adapt to changing dynamics, handle combinatorial diversity, and\nprovide heuristic search. However, limitations such as premature convergence,\nlack of problem-specific knowledge, and randomness of crossover and mutation\noperators make GAs generally inefficient in finding an optimal solution. To\naddress these limitations, this paper proposes a new metaheuristic algorithm\ncalled the Genetic Engineering Algorithm (GEA) that draws inspiration from\ngenetic engineering concepts. GEA redesigns the traditional GA while\nincorporating new search methods to isolate, purify, insert, and express new\ngenes based on existing ones, leading to the emergence of desired traits and\nthe production of specific chromosomes based on the selected genes. Comparative\nevaluations against state-of-the-art algorithms on benchmark instances\ndemonstrate the superior performance of GEA, showcasing its potential as an\ninnovative and efficient solution for combinatorial optimization problems.\n", "date": "2023-09-28", "categories": []} {"id": "2309.15286", "title": "Composable Coresets for Determinant Maximization: Greedy is Almost\n Optimal", "abstract": " Given a set of $n$ vectors in $\\mathbb{R}^d$, the goal of the\n\\emph{determinant maximization} problem is to pick $k$ vectors with the maximum\nvolume. Determinant maximization is the MAP-inference task for determinantal\npoint processes (DPP) and has recently received considerable attention for\nmodeling diversity. As most applications for the problem use large amounts of\ndata, this problem has been studied in the relevant \\textit{composable coreset}\nsetting. In particular, [Indyk-Mahabadi-OveisGharan-Rezaei--SODA'20, ICML'19]\nshowed that one can get composable coresets with optimal approximation factor\nof $\\tilde O(k)^k$ for the problem, and that a local search algorithm achieves\nan almost optimal approximation guarantee of $O(k)^{2k}$. In this work, we show\nthat the widely-used Greedy algorithm also provides composable coresets with an\nalmost optimal approximation factor of $O(k)^{3k}$, which improves over the\npreviously known guarantee of $C^{k^2}$, and supports the prior experimental\nresults showing the practicality of the greedy algorithm as a coreset. Our main\nresult follows by showing a local optimality property for Greedy: swapping a\nsingle point from the greedy solution with a vector that was not picked by the\ngreedy algorithm can increase the volume by a factor of at most $(1+\\sqrt{k})$.\nThis is tight up to the additive constant $1$. Finally, our experiments show\nthat the local optimality of the greedy algorithm is even lower than the\ntheoretical bound on real data sets.\n", "date": "2023-09-26", "categories": []} {"id": "2309.14359", "title": "Optimizing Chance-Constrained Submodular Problems with Variable\n Uncertainties", "abstract": " Chance constraints are frequently used to limit the probability of constraint\nviolations in real-world optimization problems where the constraints involve\nstochastic components. We study chance-constrained submodular optimization\nproblems, which capture a wide range of optimization problems with stochastic\nconstraints. Previous studies considered submodular problems with stochastic\nknapsack constraints in the case where uncertainties are the same for each item\nthat can be selected. However, uncertainty levels are usually variable with\nrespect to the different stochastic components in real-world scenarios, and\nrigorous analysis for this setting is missing in the context of submodular\noptimization. This paper provides the first such analysis for this case, where\nthe weights of items have the same expectation but different dispersion. We\npresent greedy algorithms that can obtain a high-quality solution, i.e., a\nconstant approximation ratio to the given optimal solution from the\ndeterministic setting. In the experiments, we demonstrate that the algorithms\nperform effectively on several chance-constrained instances of the maximum\ncoverage problem and the influence maximization problem.\n", "date": "2023-09-23", "categories": []} {"id": "2303.01543", "title": "Decision-Oriented Learning with Differentiable Submodular Maximization\n for Vehicle Routing Problem", "abstract": " We study the problem of learning a function that maps context observations\n(input) to parameters of a submodular function (output). Our motivating case\nstudy is a specific type of vehicle routing problem, in which a team of\nUnmanned Ground Vehicles (UGVs) can serve as mobile charging stations to\nrecharge a team of Unmanned Ground Vehicles (UAVs) that execute persistent\nmonitoring tasks. {We want to learn the mapping from observations of UAV task\nroutes and wind field to the parameters of a submodular objective function,\nwhich describes the distribution of landing positions of the UAVs .}\nTraditionally, such a learning problem is solved independently as a prediction\nphase without considering the downstream task optimization phase. However, the\nloss function used in prediction may be misaligned with our final goal, i.e., a\ngood routing decision. Good performance in the isolated prediction phase does\nnot necessarily lead to good decisions in the downstream routing task. In this\npaper, we propose a framework that incorporates task optimization as a\ndifferentiable layer in the prediction phase. Our framework allows end-to-end\ntraining of the prediction model without using engineered intermediate loss\nthat is targeted only at the prediction performance. In the proposed framework,\ntask optimization (submodular maximization) is made differentiable by\nintroducing stochastic perturbations into deterministic algorithms (i.e.,\nstochastic smoothing). We demonstrate the efficacy of the proposed framework\nusing synthetic data. Experimental results of the mobile charging station\nrouting problem show that the proposed framework can result in better routing\ndecisions, e.g. the average number of UAVs recharged increases, compared to the\nprediction-optimization separate approach.\n", "date": "2023-03-02", "categories": []} {"id": "2302.10848", "title": "Deep reinforced learning heuristic tested on spin-glass ground states:\n The larger picture", "abstract": " In Changjun Fan et al. [Nature Communications\nhttps://doi.org/10.1038/s41467-023-36363-w (2023)], the authors present a deep\nreinforced learning approach to augment combinatorial optimization heuristics.\nIn particular, they present results for several spin glass ground state\nproblems, for which instances on non-planar networks are generally NP-hard, in\ncomparison with several Monte Carlo based methods, such as simulated annealing\n(SA) or parallel tempering (PT). Indeed, those results demonstrate that the\nreinforced learning improves the results over those obtained with SA or PT, or\nat least allows for reduced runtimes for the heuristics before results of\ncomparable quality have been obtained relative to those other methods. To\nfacilitate the conclusion that their method is ''superior'', the authors pursue\ntwo basic strategies: (1) A commercial GUROBI solver is called on to procure a\nsample of exact ground states as a testbed to compare with, and (2) a\nhead-to-head comparison between the heuristics is given for a sample of larger\ninstances where exact ground states are hard to ascertain. Here, we put these\nstudies into a larger context, showing that the claimed superiority is at best\nmarginal for smaller samples and becomes essentially irrelevant with respect to\nany sensible approximation of true ground states in the larger samples. For\nexample, this method becomes irrelevant as a means to determine stiffness\nexponents $\\theta$ in $d>2$, as mentioned by the authors, where the problem is\nnot only NP-hard but requires the subtraction of two almost equal ground-state\nenergies and systemic errors in each of $\\approx 1\\%$ found here are\nunacceptable. This larger picture on the method arises from a straightforward\nfinite-size corrections study over the spin glass ensembles the authors employ,\nusing data that has been available for decades.\n", "date": "2023-02-21", "categories": []} {"id": "2309.12025", "title": "Robust Approximation Algorithms for Non-monotone $k$-Submodular\n Maximization under a Knapsack Constraint", "abstract": " The problem of non-monotone $k$-submodular maximization under a knapsack\nconstraint ($\\kSMK$) over the ground set size $n$ has been raised in many\napplications in machine learning, such as data summarization, information\npropagation, etc. However, existing algorithms for the problem are facing\nquestioning of how to overcome the non-monotone case and how to fast return a\ngood solution in case of the big size of data. This paper introduces two\ndeterministic approximation algorithms for the problem that competitively\nimprove the query complexity of existing algorithms.\n Our first algorithm, $\\LAA$, returns an approximation ratio of $1/19$ within\n$O(nk)$ query complexity. The second one, $\\RLA$, improves the approximation\nratio to $1/5-\\epsilon$ in $O(nk)$ queries, where $\\epsilon$ is an input\nparameter.\n Our algorithms are the first ones that provide constant approximation ratios\nwithin only $O(nk)$ query complexity for the non-monotone objective. They,\ntherefore, need fewer the number of queries than state-of-the-the-art ones by a\nfactor of $\\Omega(\\log n)$.\n Besides the theoretical analysis, we have evaluated our proposed ones with\nseveral experiments in some instances: Influence Maximization and Sensor\nPlacement for the problem. The results confirm that our algorithms ensure\ntheoretical quality as the cutting-edge techniques and significantly reduce the\nnumber of queries.\n", "date": "2023-09-21", "categories": []} {"id": "2301.01323", "title": "Graphical House Allocation", "abstract": " The classical house allocation problem involves assigning $n$ houses (or\nitems) to $n$ agents according to their preferences. A key criterion in such\nproblems is satisfying some fairness constraints such as envy-freeness. We\nconsider a generalization of this problem wherein the agents are placed along\nthe vertices of a graph (corresponding to a social network), and each agent can\nonly experience envy towards its neighbors. Our goal is to minimize the\naggregate envy among the agents as a natural fairness objective, i.e., the sum\nof all pairwise envy values over all edges in a social graph.\n When agents have identical and evenly-spaced valuations, our problem reduces\nto the well-studied problem of linear arrangements. For identical valuations\nwith possibly uneven spacing, we show a number of deep and surprising ways in\nwhich our setting is a departure from this classical problem. More broadly, we\ncontribute several structural and computational results for various classes of\ngraphs, including NP-hardness results for disjoint unions of paths, cycles,\nstars, or cliques, and fixed-parameter tractable (and, in some cases,\npolynomial-time) algorithms for paths, cycles, stars, cliques, and their\ndisjoint unions. Additionally, a conceptual contribution of our work is the\nformulation of a structural property for disconnected graphs that we call\nseparability which results in efficient parameterized algorithms for finding\noptimal allocations.\n", "date": "2023-01-03", "categories": []} {"id": "2307.14799", "title": "Hybrid ASP-based multi-objective scheduling of semiconductor\n manufacturing processes (Extended version)", "abstract": " Modern semiconductor manufacturing involves intricate production processes\nconsisting of hundreds of operations, which can take several months from lot\nrelease to completion. The high-tech machines used in these processes are\ndiverse, operate on individual wafers, lots, or batches in multiple stages, and\nnecessitate product-specific setups and specialized maintenance procedures.\nThis situation is different from traditional job-shop scheduling scenarios,\nwhich have less complex production processes and machines, and mainly focus on\nsolving highly combinatorial but abstract scheduling problems. In this work, we\naddress the scheduling of realistic semiconductor manufacturing processes by\nmodeling their specific requirements using hybrid Answer Set Programming with\ndifference logic, incorporating flexible machine processing, setup, batching\nand maintenance operations. Unlike existing methods that schedule semiconductor\nmanufacturing processes locally with greedy heuristics or by independently\noptimizing specific machine group allocations, we examine the potentials of\nlarge-scale scheduling subject to multiple optimization objectives.\n", "date": "2023-07-27", "categories": []} {"id": "2308.07347", "title": "A Parallel Ensemble of Metaheuristic Solvers for the Traveling Salesman\n Problem", "abstract": " The travelling salesman problem (TSP) is one of the well-studied NP-hard\nproblems in the literature. The state-of-the art inexact TSP solvers are the\nLin-Kernighan-Helsgaun (LKH) heuristic and Edge Assembly crossover (EAX). A\nrecent study suggests that EAX with restart mechanisms perform well on a wide\nrange of TSP instances. However, this study is limited to 2,000 city problems.\nWe study for problems ranging from 2,000 to 85,900. We see that the performance\nof the solver varies with the type of the problem. However, combining these\nsolvers in an ensemble setup, we are able to outperform the individual solver's\nperformance. We see the ensemble setup as an efficient way to make use of the\nabundance of compute resources. In addition to EAX and LKH, we use several\nversions of the hybrid of EAX and Mixing Genetic Algorithm (MGA). A hybrid of\nMGA and EAX is known to solve some hard problems. We see that the ensemble of\nthe hybrid version outperforms the state-of-the-art solvers on problems larger\nthan 10,000 cities.\n", "date": "2023-08-13", "categories": []} {"id": "2309.06392", "title": "A Fast Algorithm for Moderating Critical Nodes via Edge Removal", "abstract": " Critical nodes in networks are extremely vulnerable to malicious attacks to\ntrigger negative cascading events such as the spread of misinformation and\ndiseases. Therefore, effective moderation of critical nodes is very vital for\nmitigating the potential damages caused by such malicious diffusions. The\ncurrent moderation methods are computationally expensive. Furthermore, they\ndisregard the fundamental metric of information centrality, which measures the\ndissemination power of nodes.\n We investigate the problem of removing $k$ edges from a network to minimize\nthe information centrality of a target node $\\lea$ while preserving the\nnetwork's connectivity. We prove that this problem is computationally\nchallenging: it is NP-complete and its objective function is not supermodular.\nHowever, we propose three approximation greedy algorithms using novel\ntechniques such as random walk-based Schur complement approximation and fast\nsum estimation. One of our algorithms runs in nearly linear time in the number\nof edges.\n To complement our theoretical analysis, we conduct a comprehensive set of\nexperiments on synthetic and real networks with over one million nodes. Across\nvarious settings, the experimental results illustrate the effectiveness and\nefficiency of our proposed algorithms.\n", "date": "2023-09-09", "categories": []} {"id": "2210.11321", "title": "A Study of Scalarisation Techniques for Multi-Objective QUBO Solving", "abstract": " In recent years, there has been significant research interest in solving\nQuadratic Unconstrained Binary Optimisation (QUBO) problems. Physics-inspired\noptimisation algorithms have been proposed for deriving optimal or sub-optimal\nsolutions to QUBOs. These methods are particularly attractive within the\ncontext of using specialised hardware, such as quantum computers, application\nspecific CMOS and other high performance computing resources for solving\noptimisation problems. These solvers are then applied to QUBO formulations of\ncombinatorial optimisation problems. Quantum and quantum-inspired optimisation\nalgorithms have shown promising performance when applied to academic benchmarks\nas well as real-world problems. However, QUBO solvers are single objective\nsolvers. To make them more efficient at solving problems with multiple\nobjectives, a decision on how to convert such multi-objective problems to\nsingle-objective problems need to be made. In this study, we compare methods of\nderiving scalarisation weights when combining two objectives of the cardinality\nconstrained mean-variance portfolio optimisation problem into one. We show\nsignificant performance improvement (measured in terms of hypervolume) when\nusing a method that iteratively fills the largest space in the Pareto front\ncompared to a n\\\"aive approach using uniformly generated weights.\n", "date": "2022-10-20", "categories": []} {"id": "2309.03517", "title": "Parameterized Aspects of Distinct Kemeny Rank Aggregation", "abstract": " The Kemeny method is one of the popular tools for rank aggregation. However,\ncomputing an optimal Kemeny ranking is NP-hard. Consequently, the computational\ntask of finding a Kemeny ranking has been studied under the lens of\nparameterized complexity with respect to many parameters. We first present a\ncomprehensive relationship, both theoretical and empirical, among these\nparameters. Further, we study the problem of computing all distinct Kemeny\nrankings under the lens of parameterized complexity. We consider the target\nKemeny score, number of candidates, average distance of input rankings, maximum\nrange of any candidate, and unanimity width as our parameters. For all these\nparameters, we already have FPT algorithms. We find that any desirable number\nof Kemeny rankings can also be found without substantial increase in running\ntime. We also present FPT approximation algorithms for Kemeny rank aggregation\nwith respect to these parameters.\n", "date": "2023-09-07", "categories": []} {"id": "2211.00980", "title": "Balancing Utility and Fairness in Submodular Maximization (Technical\n Report)", "abstract": " Submodular function maximization is a fundamental combinatorial optimization\nproblem with plenty of applications -- including data summarization, influence\nmaximization, and recommendation. In many of these problems, the goal is to\nfind a solution that maximizes the average utility over all users, for each of\nwhom the utility is defined by a monotone submodular function. However, when\nthe population of users is composed of several demographic groups, another\ncritical problem is whether the utility is fairly distributed across different\ngroups. Although the \\emph{utility} and \\emph{fairness} objectives are both\ndesirable, they might contradict each other, and, to the best of our knowledge,\nlittle attention has been paid to optimizing them jointly.\n To fill this gap, we propose a new problem called \\emph{Bicriteria Submodular\nMaximization} (BSM) to balance utility and fairness. Specifically, it requires\nfinding a fixed-size solution to maximize the utility function, subject to the\nvalue of the fairness function not being below a threshold. Since BSM is\ninapproximable within any constant factor, we focus on designing efficient\ninstance-dependent approximation schemes. Our algorithmic proposal comprises\ntwo methods, with different approximation factors, obtained by converting a BSM\ninstance into other submodular optimization problem instances. Using real-world\nand synthetic datasets, we showcase applications of our proposed methods in\nthree submodular maximization problems: maximum coverage, influence\nmaximization, and facility location.\n", "date": "2022-11-02", "categories": []} {"id": "2309.00197", "title": "Deep-learning-based Early Fixing for Gas-lifted Oil Production\n Optimization: Supervised and Weakly-supervised Approaches", "abstract": " Maximizing oil production from gas-lifted oil wells entails solving\nMixed-Integer Linear Programs (MILPs). As the parameters of the wells, such as\nthe basic-sediment-to-water ratio and the gas-oil ratio, are updated, the\nproblems must be repeatedly solved. Instead of relying on costly exact methods\nor the accuracy of general approximate methods, in this paper, we propose a\ntailor-made heuristic solution based on deep learning models trained to provide\nvalues to all integer variables given varying well parameters, early-fixing the\ninteger variables and, thus, reducing the original problem to a linear program\n(LP). We propose two approaches for developing the learning-based heuristic: a\nsupervised learning approach, which requires the optimal integer values for\nseveral instances of the original problem in the training set, and a\nweakly-supervised learning approach, which requires only solutions for the\nearly-fixed linear problems with random assignments for the integer variables.\nOur results show a runtime reduction of 71.11% Furthermore, the\nweakly-supervised learning model provided significant values for early fixing,\ndespite never seeing the optimal values during training.\n", "date": "2023-09-01", "categories": []} {"id": "2207.03364", "title": "Group Equality in Adaptive Submodular Maximization", "abstract": " In this paper, we study the classic submodular maximization problem subject\nto a group equality constraint under both non-adaptive and adaptive settings.\nIt has been shown that the utility function of many machine learning\napplications, including data summarization, influence maximization in social\nnetworks, and personalized recommendation, satisfies the property of\nsubmodularity. Hence, maximizing a submodular function subject to various\nconstraints can be found at the heart of many of those applications. On a high\nlevel, submodular maximization aims to select a group of most representative\nitems (e.g., data points). However, the design of most existing algorithms does\nnot incorporate the fairness constraint, leading to under- or\nover-representation of some particular groups. This motivates us to study the\nsubmodular maximization problem with group equality, where we aim to select a\ngroup of items to maximize a (possibly non-monotone) submodular utility\nfunction subject to a group equality constraint. To this end, we develop the\nfirst constant-factor approximation algorithm for this problem. The design of\nour algorithm is robust enough to be extended to solving the submodular\nmaximization problem under a more complicated adaptive setting. Moreover, we\nfurther extend our study to incorporating a global cardinality constraint and\nother fairness notations.\n", "date": "2022-07-07", "categories": []} {"id": "2308.13252", "title": "Kissing to Find a Match: Efficient Low-Rank Permutation Representation", "abstract": " Permutation matrices play a key role in matching and assignment problems\nacross the fields, especially in computer vision and robotics. However, memory\nfor explicitly representing permutation matrices grows quadratically with the\nsize of the problem, prohibiting large problem instances. In this work, we\npropose to tackle the curse of dimensionality of large permutation matrices by\napproximating them using low-rank matrix factorization, followed by a\nnonlinearity. To this end, we rely on the Kissing number theory to infer the\nminimal rank required for representing a permutation matrix of a given size,\nwhich is significantly smaller than the problem size. This leads to a drastic\nreduction in computation and memory costs, e.g., up to $3$ orders of magnitude\nless memory for a problem of size $n=20000$, represented using $8.4\\times10^5$\nelements in two small matrices instead of using a single huge matrix with\n$4\\times 10^8$ elements. The proposed representation allows for accurate\nrepresentations of large permutation matrices, which in turn enables handling\nlarge problems that would have been infeasible otherwise. We demonstrate the\napplicability and merits of the proposed approach through a series of\nexperiments on a range of problems that involve predicting permutation\nmatrices, from linear and quadratic assignment to shape matching problems.\n", "date": "2023-08-25", "categories": []} {"id": "2308.12606", "title": "A Greedy Approach for Offering to Telecom Subscribers", "abstract": " Customer retention or churn prevention is a challenging task of a telecom\noperator. One of the effective approaches is to offer some attractive incentive\nor additional services or money to the subscribers for keeping them engaged and\nmake sure they stay in the operator's network for longer time. Often, operators\nallocate certain amount of monetary budget to carry out the offer campaign. The\ndifficult part of this campaign is the selection of a set of customers from a\nlarge subscriber-base and deciding the amount that should be offered to an\nindividual so that operator's objective is achieved. There may be multiple\nobjectives (e.g., maximizing revenue, minimizing number of churns) for\nselection of subscriber and selection of an offer to the selected subscriber.\nApart from monetary benefit, offers may include additional data, SMS, hots-spot\ntethering, and many more. This problem is known as offer optimization. In this\npaper, we propose a novel combinatorial algorithm for solving offer\noptimization under heterogeneous offers by maximizing expected revenue under\nthe scenario of subscriber churn, which is, in general, seen in telecom domain.\nThe proposed algorithm is efficient and accurate even for a very large\nsubscriber-base.\n", "date": "2023-08-24", "categories": []} {"id": "2308.06644", "title": "Accelerating Diffusion-based Combinatorial Optimization Solvers by\n Progressive Distillation", "abstract": " Graph-based diffusion models have shown promising results in terms of\ngenerating high-quality solutions to NP-complete (NPC) combinatorial\noptimization (CO) problems. However, those models are often inefficient in\ninference, due to the iterative evaluation nature of the denoising diffusion\nprocess. This paper proposes to use progressive distillation to speed up the\ninference by taking fewer steps (e.g., forecasting two steps ahead within a\nsingle step) during the denoising process. Our experimental results show that\nthe progressively distilled model can perform inference 16 times faster with\nonly 0.019% degradation in performance on the TSP-50 dataset.\n", "date": "2023-08-12", "categories": []} {"id": "2308.11652", "title": "Accelerating Exact Combinatorial Optimization via RL-based\n Initialization -- A Case Study in Scheduling", "abstract": " Scheduling on dataflow graphs (also known as computation graphs) is an\nNP-hard problem. The traditional exact methods are limited by runtime\ncomplexity, while reinforcement learning (RL) and heuristic-based approaches\nstruggle with determinism and solution quality. This research aims to develop\nan innovative approach that employs machine learning (ML) for addressing\ncombinatorial optimization problems, using scheduling as a case study. The goal\nis to provide guarantees in optimality and determinism while maintaining the\nruntime cost of heuristic methods. Specifically, we introduce a novel two-phase\nRL-to-ILP scheduling framework, which includes three steps: 1) RL solver acts\nas coarse-grain scheduler, 2) solution relaxation and 3) exact solving via ILP.\nOur framework demonstrates the same scheduling performance compared with using\nexact scheduling methods while achieving up to 128 $\\times$ speed improvements.\nThis was conducted on actual EdgeTPU platforms, utilizing ImageNet DNN\ncomputation graphs as input. Additionally, the framework offers improved\non-chip inference runtime and acceleration compared to the commercially\navailable EdgeTPU compiler.\n", "date": "2023-08-19", "categories": []} {"id": "2308.11943", "title": "RamseyRL: A Framework for Intelligent Ramsey Number Counterexample\n Searching", "abstract": " The Ramsey number is the minimum number of nodes, $n = R(s, t)$, such that\nall undirected simple graphs of order $n$, contain a clique of order $s$, or an\nindependent set of order $t$. This paper explores the application of a best\nfirst search algorithm and reinforcement learning (RL) techniques to find\ncounterexamples to specific Ramsey numbers. We incrementally improve over prior\nsearch methods such as random search by introducing a graph vectorization and\ndeep neural network (DNN)-based heuristic, which gauge the likelihood of a\ngraph being a counterexample. The paper also proposes algorithmic optimizations\nto confine a polynomial search runtime. This paper does not aim to present new\ncounterexamples but rather introduces and evaluates a framework supporting\nRamsey counterexample exploration using other heuristics. Code and methods are\nmade available through a PyPI package and GitHub repository.\n", "date": "2023-08-23", "categories": []} {"id": "2305.11195", "title": "DClEVerNet: Deep Combinatorial Learning for Efficient EV Charging\n Scheduling in Large-scale Networked Facilities", "abstract": " With the electrification of transportation, the rising uptake of electric\nvehicles (EVs) might stress distribution networks significantly, leaving their\nperformance degraded and stability jeopardized. To accommodate these new loads\ncost-effectively, modern power grids require coordinated or ``smart'' charging\nstrategies capable of optimizing EV charging scheduling in a scalable and\nefficient fashion. With this in view, the present work focuses on reservation\nmanagement programs for large-scale, networked EV charging stations. We\nformulate a time-coupled binary optimization problem that maximizes EV users'\ntotal welfare gain while accounting for the network's available power capacity\nand stations' occupancy limits. To tackle the problem at scale while retaining\nhigh solution quality, a data-driven optimization framework combining\ntechniques from the fields of Deep Learning and Approximation Algorithms is\nintroduced. The framework's key ingredient is a novel input-output processing\nscheme for neural networks that allows direct extrapolation to problem sizes\nsubstantially larger than those included in the training set. Extensive\nnumerical simulations based on synthetic and real-world data traces verify the\neffectiveness and superiority of the presented approach over two representative\nscheduling algorithms. Lastly, we round up the contributions by listing several\nimmediate extensions to the proposed framework and outlining the prospects for\nfurther exploration.\n", "date": "2023-05-18", "categories": []} {"id": "2207.10240", "title": "Differentially Private Partial Set Cover with Applications to Facility\n Location", "abstract": " It was observed in \\citet{gupta2009differentially} that the Set Cover problem\nhas strong impossibility results under differential privacy. In our work, we\nobserve that these hardness results dissolve when we turn to the Partial Set\nCover problem, where we only need to cover a $\\rho$-fraction of the elements in\nthe universe, for some $\\rho\\in(0,1)$. We show that this relaxation enables us\nto avoid the impossibility results: under loose conditions on the input set\nsystem, we give differentially private algorithms which output an explicit set\ncover with non-trivial approximation guarantees. In particular, this is the\nfirst differentially private algorithm which outputs an explicit set cover.\n Using our algorithm for Partial Set Cover as a subroutine, we give a\ndifferentially private (bicriteria) approximation algorithm for a facility\nlocation problem which generalizes $k$-center/$k$-supplier with outliers. Like\nwith the Set Cover problem, no algorithm has been able to give non-trivial\nguarantees for $k$-center/$k$-supplier-type facility location problems due to\nthe high sensitivity and impossibility results. Our algorithm shows that\nrelaxing the covering requirement to serving only a $\\rho$-fraction of the\npopulation, for $\\rho\\in(0,1)$, enables us to circumvent the inherent hardness.\nOverall, our work is an important step in tackling and understanding\nimpossibility results in private combinatorial optimization.\n", "date": "2022-07-21", "categories": []} {"id": "2305.13459", "title": "The First Proven Performance Guarantees for the Non-Dominated Sorting\n Genetic Algorithm II (NSGA-II) on a Combinatorial Optimization Problem", "abstract": " The Non-dominated Sorting Genetic Algorithm-II (NSGA-II) is one of the most\nprominent algorithms to solve multi-objective optimization problems. Recently,\nthe first mathematical runtime guarantees have been obtained for this\nalgorithm, however only for synthetic benchmark problems.\n In this work, we give the first proven performance guarantees for a classic\noptimization problem, the NP-complete bi-objective minimum spanning tree\nproblem. More specifically, we show that the NSGA-II with population size $N\n\\ge 4((n-1) w_{\\max} + 1)$ computes all extremal points of the Pareto front in\nan expected number of $O(m^2 n w_{\\max} \\log(n w_{\\max}))$ iterations, where\n$n$ is the number of vertices, $m$ the number of edges, and $w_{\\max}$ is the\nmaximum edge weight in the problem instance. This result confirms, via\nmathematical means, the good performance of the NSGA-II observed empirically.\nIt also shows that mathematical analyses of this algorithm are not only\npossible for synthetic benchmark problems, but also for more complex\ncombinatorial optimization problems.\n As a side result, we also obtain a new analysis of the performance of the\nglobal SEMO algorithm on the bi-objective minimum spanning tree problem, which\nimproves the previous best result by a factor of $|F|$, the number of extremal\npoints of the Pareto front, a set that can be as large as $n w_{\\max}$. The\nmain reason for this improvement is our observation that both multi-objective\nevolutionary algorithms find the different extremal points in parallel rather\nthan sequentially, as assumed in the previous proofs.\n", "date": "2023-05-22", "categories": []} {"id": "2304.11726", "title": "End-to-End Feasible Optimization Proxies for Large-Scale Economic\n Dispatch", "abstract": " The paper proposes a novel End-to-End Learning and Repair (E2ELR)\narchitecture for training optimization proxies for economic dispatch problems.\nE2ELR combines deep neural networks with closed-form, differentiable repair\nlayers, thereby integrating learning and feasibility in an end-to-end fashion.\nE2ELR is also trained with self-supervised learning, removing the need for\nlabeled data and the solving of numerous optimization problems offline. E2ELR\nis evaluated on industry-size power grids with tens of thousands of buses using\nan economic dispatch that co-optimizes energy and reserves. The results\ndemonstrate that the self-supervised E2ELR achieves state-of-the-art\nperformance, with optimality gaps that outperform other baselines by at least\nan order of magnitude.\n", "date": "2023-04-23", "categories": []} {"id": "2308.07335", "title": "An Encoder-Decoder Approach for Packing Circles", "abstract": " The problem of packing smaller objects within a larger object has been of\ninterest since decades. In these problems, in addition to the requirement that\nthe smaller objects must lie completely inside the larger objects, they are\nexpected to not overlap or have minimum overlap with each other. Due to this,\nthe problem of packing turns out to be a non-convex problem, obtaining whose\noptimal solution is challenging. As such, several heuristic approaches have\nbeen used for obtaining sub-optimal solutions in general, and provably optimal\nsolutions for some special instances. In this paper, we propose a novel\nencoder-decoder architecture consisting of an encoder block, a perturbation\nblock and a decoder block, for packing identical circles within a larger\ncircle. In our approach, the encoder takes the index of a circle to be packed\nas an input and outputs its center through a normalization layer, the\nperturbation layer adds controlled perturbations to the center, ensuring that\nit does not deviate beyond the radius of the smaller circle to be packed, and\nthe decoder takes the perturbed center as input and estimates the index of the\nintended circle for packing. We parameterize the encoder and decoder by a\nneural network and optimize it to reduce an error between the decoder's\nestimated index and the actual index of the circle provided as input to the\nencoder. The proposed approach can be generalized to pack objects of higher\ndimensions and different shapes by carefully choosing normalization and\nperturbation layers. The approach gives a sub-optimal solution and is able to\npack smaller objects within a larger object with competitive performance with\nrespect to classical methods.\n", "date": "2023-08-11", "categories": []} {"id": "2306.17609", "title": "Mixed Integer Programming for Time-Optimal Multi-Robot Coverage Path\n Planning with Efficient Heuristics", "abstract": " We investigate time-optimal Multi-Robot Coverage Path Planning (MCPP) for\nboth unweighted and weighted terrains, which aims to minimize the coverage\ntime, defined as the maximum travel time of all robots. Specifically, we focus\non a reduction from MCPP to Min-Max Rooted Tree Cover (MMRTC). For the first\ntime, we propose a Mixed Integer Programming (MIP) model to optimally solve\nMMRTC, resulting in an MCPP solution with a coverage time that is provably at\nmost four times the optimal. Moreover, we propose two suboptimal yet effective\nheuristics that reduce the number of variables in the MIP model, thus improving\nits efficiency for large-scale MCPP instances. We show that both heuristics\nresult in reduced-size MIP models that remain complete (i.e., guaranteed to\nfind a solution if one exists) for all MMRTC instances. Additionally, we\nexplore the use of model optimization warm-startup to further improve the\nefficiency of both the original MIP model and the reduced-size MIP models. We\nvalidate the effectiveness of our MIP-based MCPP planner through experiments\nthat compare it with two state-of-the-art MCPP planners on various instances,\ndemonstrating a reduction in the coverage time by an average of 27.65% and\n23.24% over them, respectively.\n", "date": "2023-06-30", "categories": []} {"id": "2308.04639", "title": "A Hierarchical Destroy and Repair Approach for Solving Very Large-Scale\n Travelling Salesman Problem", "abstract": " For prohibitively large-scale Travelling Salesman Problems (TSPs), existing\nalgorithms face big challenges in terms of both computational efficiency and\nsolution quality. To address this issue, we propose a hierarchical\ndestroy-and-repair (HDR) approach, which attempts to improve an initial\nsolution by applying a series of carefully designed destroy-and-repair\noperations. A key innovative concept is the hierarchical search framework,\nwhich recursively fixes partial edges and compresses the input instance into a\nsmall-scale TSP under some equivalence guarantee. This neat search framework is\nable to deliver highly competitive solutions within a reasonable time. Fair\ncomparisons based on nineteen famous large-scale instances (with 10,000 to\n10,000,000 cities) show that HDR is highly competitive against existing\nstate-of-the-art TSP algorithms, in terms of both efficiency and solution\nquality. Notably, on two large instances with 3,162,278 and 10,000,000 cities,\nHDR breaks the world records (i.e., best-known results regardless of\ncomputation time), which were previously achieved by LKH and its variants,\nwhile HDR is completely independent of LKH. Finally, ablation studies are\nperformed to certify the importance and validity of the hierarchical search\nframework.\n", "date": "2023-08-09", "categories": []} {"id": "2212.11372", "title": "GCS-Q: Quantum Graph Coalition Structure Generation", "abstract": " The problem of generating an optimal coalition structure for a given\ncoalition game of rational agents is to find a partition that maximizes their\nsocial welfare and is known to be NP-hard. This paper proposes GCS-Q, a novel\nquantum-supported solution for Induced Subgraph Games (ISGs) in coalition\nstructure generation. GCS-Q starts by considering the grand coalition as\ninitial coalition structure and proceeds by iteratively splitting the\ncoalitions into two nonempty subsets to obtain a coalition structure with a\nhigher coalition value. In particular, given an $n$-agent ISG, the GCS-Q solves\nthe optimal split problem $\\mathcal{O} (n)$ times using a quantum annealing\ndevice, exploring $\\mathcal{O}(2^n)$ partitions at each step. We show that\nGCS-Q outperforms the currently best classical solvers with its runtime in the\norder of $n^2$ and an expected worst-case approximation ratio of $93\\%$ on\nstandard benchmark datasets.\n", "date": "2022-12-21", "categories": []} {"id": "2302.07181", "title": "Quantum algorithms applied to satellite mission planning for Earth\n observation", "abstract": " Earth imaging satellites are a crucial part of our everyday lives that enable\nglobal tracking of industrial activities. Use cases span many applications,\nfrom weather forecasting to digital maps, carbon footprint tracking, and\nvegetation monitoring. However, there are limitations; satellites are difficult\nto manufacture, expensive to maintain, and tricky to launch into orbit.\nTherefore, satellites must be employed efficiently. This poses a challenge\nknown as the satellite mission planning problem, which could be computationally\nprohibitive to solve on large scales. However, close-to-optimal algorithms,\nsuch as greedy reinforcement learning and optimization algorithms, can often\nprovide satisfactory resolutions. This paper introduces a set of quantum\nalgorithms to solve the mission planning problem and demonstrate an advantage\nover the classical algorithms implemented thus far. The problem is formulated\nas maximizing the number of high-priority tasks completed on real datasets\ncontaining thousands of tasks and multiple satellites. This work demonstrates\nthat through solution-chaining and clustering, optimization and machine\nlearning algorithms offer the greatest potential for optimal solutions. This\npaper notably illustrates that a hybridized quantum-enhanced reinforcement\nlearning agent can achieve a completion percentage of 98.5% over high-priority\ntasks, significantly improving over the baseline greedy methods with a\ncompletion rate of 75.8%. The results presented in this work pave the way to\nquantum-enabled solutions in the space industry and, more generally, future\nmission planning problems across industries.\n", "date": "2023-02-14", "categories": []} {"id": "2308.03185", "title": "VN-Solver: Vision-based Neural Solver for Combinatorial Optimization\n over Graphs", "abstract": " Data-driven approaches have been proven effective in solving combinatorial\noptimization problems over graphs such as the traveling salesman problems and\nthe vehicle routing problem. The rationale behind such methods is that the\ninput instances may follow distributions with salient patterns that can be\nleveraged to overcome the worst-case computational hardness. For optimization\nproblems over graphs, the common practice of neural combinatorial solvers\nconsumes the inputs in the form of adjacency matrices. In this paper, we\nexplore a vision-based method that is conceptually novel: can neural models\nsolve graph optimization problems by \\textit{taking a look at the graph\npattern}? Our results suggest that the performance of such vision-based methods\nis not only non-trivial but also comparable to the state-of-the-art\nmatrix-based methods, which opens a new avenue for developing data-driven\noptimization solvers.\n", "date": "2023-08-06", "categories": []} {"id": "2303.01977", "title": "Hybrid Approach for Solving Real-World Bin Packing Problem Instances\n Using Quantum Annealers", "abstract": " Efficient packing of items into bins is a common daily task. Known as Bin\nPacking Problem, it has been intensively studied in the field of artificial\nintelligence, thanks to the wide interest from industry and logistics. Since\ndecades, many variants have been proposed, with the three-dimensional Bin\nPacking Problem as the closest one to real-world use cases. We introduce a\nhybrid quantum-classical framework for solving real-world three-dimensional Bin\nPacking Problems (Q4RealBPP), considering different realistic characteristics,\nsuch as: i) package and bin dimensions, ii) overweight restrictions, iii)\naffinities among item categories and iv) preferences for item ordering.\nQ4RealBPP permits the solving of real-world oriented instances of 3dBPP,\ncontemplating restrictions well appreciated by industrial and logistics\nsectors.\n", "date": "2023-03-01", "categories": []} {"id": "2308.01797", "title": "Job Shop Scheduling via Deep Reinforcement Learning: a Sequence to\n Sequence approach", "abstract": " Job scheduling is a well-known Combinatorial Optimization problem with\nendless applications. Well planned schedules bring many benefits in the context\nof automated systems: among others, they limit production costs and waste.\nNevertheless, the NP-hardness of this problem makes it essential to use\nheuristics whose design is difficult, requires specialized knowledge and often\nproduces methods tailored to the specific task. This paper presents an original\nend-to-end Deep Reinforcement Learning approach to scheduling that\nautomatically learns dispatching rules. Our technique is inspired by natural\nlanguage encoder-decoder models for sequence processing and has never been\nused, to the best of our knowledge, for scheduling purposes. We applied and\ntested our method in particular to some benchmark instances of Job Shop\nProblem, but this technique is general enough to be potentially used to tackle\nother different optimal job scheduling tasks with minimal intervention. Results\ndemonstrate that we outperform many classical approaches exploiting priority\ndispatching rules and show competitive results on state-of-the-art Deep\nReinforcement Learning ones.\n", "date": "2023-08-03", "categories": []} {"id": "2308.00327", "title": "Threshold-aware Learning to Generate Feasible Solutions for Mixed\n Integer Programs", "abstract": " Finding a high-quality feasible solution to a combinatorial optimization (CO)\nproblem in a limited time is challenging due to its discrete nature. Recently,\nthere has been an increasing number of machine learning (ML) methods for\naddressing CO problems. Neural diving (ND) is one of the learning-based\napproaches to generating partial discrete variable assignments in Mixed Integer\nPrograms (MIP), a framework for modeling CO problems. However, a major drawback\nof ND is a large discrepancy between the ML and MIP objectives, i.e., variable\nvalue classification accuracy over primal bound. Our study investigates that a\nspecific range of variable assignment rates (coverage) yields high-quality\nfeasible solutions, where we suggest optimizing the coverage bridges the gap\nbetween the learning and MIP objectives. Consequently, we introduce a post-hoc\nmethod and a learning-based approach for optimizing the coverage. A key idea of\nour approach is to jointly learn to restrict the coverage search space and to\npredict the coverage in the learned search space. Experimental results\ndemonstrate that learning a deep neural network to estimate the coverage for\nfinding high-quality feasible solutions achieves state-of-the-art performance\nin NeurIPS ML4CO datasets. In particular, our method shows outstanding\nperformance in the workload apportionment dataset, achieving the optimality gap\nof 0.45%, a ten-fold improvement over SCIP within the one-minute time limit.\n", "date": "2023-08-01", "categories": []} {"id": "2307.16815", "title": "A Dual-mode Local Search Algorithm for Solving the Minimum Dominating\n Set Problem", "abstract": " Given a graph, the minimum dominating set (MinDS) problem is to identify a\nsmallest set $D$ of vertices such that every vertex not in $D$ is adjacent to\nat least one vertex in $D$. The MinDS problem is a classic $\\mathcal{NP}$-hard\nproblem and has been extensively studied because of its many disparate\napplications in network analysis. To solve this problem efficiently, many\nheuristic approaches have been proposed to obtain a good solution within an\nacceptable time limit. However, existing MinDS heuristic algorithms are always\nlimited by various tie-breaking cases when selecting vertices, which slows down\nthe effectiveness of the algorithms. In this paper, we design an efficient\nlocal search algorithm for the MinDS problem, named DmDS -- a dual-mode local\nsearch framework that probabilistically chooses between two distinct\nvertex-swapping schemes. We further address limitations of other algorithms by\nintroducing vertex selection criterion based on the frequency of vertices added\nto solutions to address tie-breaking cases, and a new strategy to improve the\nquality of the initial solution via a greedy-based strategy integrated with\nperturbation. We evaluate DmDS against the state-of-the-art algorithms on seven\ndatasets, consisting of 346 instances (or families) with up to tens of millions\nof vertices. Experimental results show that DmDS obtains the best performance\nin accuracy for almost all instances and finds much better solutions than\nstate-of-the-art MinDS algorithms on a broad range of large real-world graphs.\n", "date": "2023-07-25", "categories": []} {"id": "2307.13470", "title": "Combinatorial Auctions and Graph Neural Networks for Local Energy\n Flexibility Markets", "abstract": " This paper proposes a new combinatorial auction framework for local energy\nflexibility markets, which addresses the issue of prosumers' inability to\nbundle multiple flexibility time intervals. To solve the underlying NP-complete\nwinner determination problems, we present a simple yet powerful heterogeneous\ntri-partite graph representation and design graph neural network-based models.\nOur models achieve an average optimal value deviation of less than 5\\% from an\noff-the-shelf optimization tool and show linear inference time complexity\ncompared to the exponential complexity of the commercial solver. Contributions\nand results demonstrate the potential of using machine learning to efficiently\nallocate energy flexibility resources in local markets and solving optimization\nproblems in general.\n", "date": "2023-07-25", "categories": []} {"id": "2307.13996", "title": "Fast algorithms for k-submodular maximization subject to a matroid\n constraint", "abstract": " In this paper, we apply a Threshold-Decreasing Algorithm to maximize\n$k$-submodular functions under a matroid constraint, which reduces the query\ncomplexity of the algorithm compared to the greedy algorithm with little loss\nin approximation ratio. We give a $(\\frac{1}{2} - \\epsilon)$-approximation\nalgorithm for monotone $k$-submodular function maximization, and a\n$(\\frac{1}{3} - \\epsilon)$-approximation algorithm for non-monotone case, with\ncomplexity $O(\\frac{n(k\\cdot EO + IO)}{\\epsilon} \\log \\frac{r}{\\epsilon})$,\nwhere $r$ denotes the rank of the matroid, and $IO, EO$ denote the number of\noracles to evaluate whether a subset is an independent set and to compute the\nfunction value of $f$, respectively. Since the constraint of total size can be\nlooked as a special matroid, called uniform matroid, then we present the fast\nalgorithm for maximizing $k$-submodular functions subject to a total size\nconstraint as corollaries. corollaries.\n", "date": "2023-07-26", "categories": []} {"id": "2307.13453", "title": "Monte-Carlo Tree Search for Multi-Agent Pathfinding: Preliminary Results", "abstract": " In this work we study a well-known and challenging problem of Multi-agent\nPathfinding, when a set of agents is confined to a graph, each agent is\nassigned a unique start and goal vertices and the task is to find a set of\ncollision-free paths (one for each agent) such that each agent reaches its\nrespective goal. We investigate how to utilize Monte-Carlo Tree Search (MCTS)\nto solve the problem. Although MCTS was shown to demonstrate superior\nperformance in a wide range of problems like playing antagonistic games (e.g.\nGo, Chess etc.), discovering faster matrix multiplication algorithms etc., its\napplication to the problem at hand was not well studied before. To this end we\nintroduce an original variant of MCTS, tailored to multi-agent pathfinding. The\ncrux of our approach is how the reward, that guides MCTS, is computed.\nSpecifically, we use individual paths to assist the agents with the the\ngoal-reaching behavior, while leaving them freedom to get off the track if it\nis needed to avoid collisions. We also use a dedicated decomposition technique\nto reduce the branching factor of the tree search procedure. Empirically we\nshow that the suggested method outperforms the baseline planning algorithm that\ninvokes heuristic search, e.g. A*, at each re-planning step.\n", "date": "2023-07-25", "categories": []} {"id": "2212.02230", "title": "A Hybrid Evolutionary Approach to Solve University Course Allocation\n Problem", "abstract": " This paper discusses various types of constraints, difficulties and solutions\nto overcome the challenges regarding university course allocation problem. A\nhybrid evolutionary algorithm has been defined combining Local Repair Algorithm\nand Modified Genetic Algorithm to generate the best course assignment. After\nanalyzing the collected dataset, all the necessary constraints were formulated.\nThese constraints manage to cover the aspects needed to be kept in mind while\npreparing clash free and efficient class schedules for every faculty member.\nThe goal is to generate an optimized solution which will fulfill those\nconstraints while maintaining time efficiency and also reduce the workload of\nhandling this task manually. The proposed algorithm was compared with some base\nlevel optimization algorithms to show the better efficiency in terms of\naccuracy and time.\n", "date": "2022-11-15", "categories": []} {"id": "2304.13671", "title": "Multiobjective Logistics Optimization for Automated ATM Cash\n Replenishment Process", "abstract": " In the digital transformation era, integrating digital technology into every\naspect of banking operations improves process automation, cost efficiency, and\nservice level improvement. Although logistics for ATM cash is a crucial task\nthat impacts operating costs and consumer satisfaction, there has been little\neffort to enhance it. Specifically, in Vietnam, with a market of more than\n20,000 ATMs nationally, research and technological solutions that can resolve\nthis issue remain scarce. In this paper, we generalized the vehicle routing\nproblem for ATM cash replenishment, suggested a mathematical model and then\noffered a tool to evaluate various situations. When being evaluated on the\nsimulated dataset, our proposed model and method produced encouraging results\nwith the benefits of cutting ATM cash operating costs.\n", "date": "2023-04-23", "categories": []} {"id": "2307.12133", "title": "Route Planning Using Nature-Inspired Algorithms", "abstract": " There are many different heuristic algorithms for solving combinatorial\noptimization problems that are commonly described as Nature-Inspired Algorithms\n(NIAs). Generally, they are inspired by some natural phenomenon, and due to\ntheir inherent converging and stochastic nature, they are known to give optimal\nresults when compared to classical approaches. There are a large number of\napplications of NIAs, perhaps the most popular being route planning problems in\nrobotics - problems that require a sequence of translation and rotation steps\nfrom the start to the goal in an optimized manner while avoiding obstacles in\nthe environment. In this chapter, we will first give an overview of\nNature-Inspired Algorithms, followed by their classification and common\nexamples. We will then discuss how the NIAs have applied to solve the route\nplanning problem.\n", "date": "2023-07-22", "categories": []} {"id": "2307.11621", "title": "On the Complexity of the Bipartite Polarization Problem: from Neutral to\n Highly Polarized Discussions", "abstract": " The Bipartite Polarization Problem is an optimization problem where the goal\nis to find the highest polarized bipartition on a weighted and labelled graph\nthat represents a debate developed through some social network, where nodes\nrepresent user's opinions and edges agreement or disagreement between users.\nThis problem can be seen as a generalization of the maxcut problem, and in\nprevious work approximate solutions and exact solutions have been obtained for\nreal instances obtained from Reddit discussions, showing that such real\ninstances seem to be very easy to solve. In this paper, we investigate further\nthe complexity of this problem, by introducing an instance generation model\nwhere a single parameter controls the polarization of the instances in such a\nway that this correlates with the average complexity to solve those instances.\nThe average complexity results we obtain are consistent with our hypothesis:\nthe higher the polarization of the instance, the easier is to find the\ncorresponding polarized bipartition.\n", "date": "2023-07-21", "categories": []} {"id": "2307.10688", "title": "Bounded Combinatorial Reconfiguration with Answer Set Programming", "abstract": " We develop an approach called bounded combinatorial reconfiguration for\nsolving combinatorial reconfiguration problems based on Answer Set Programming\n(ASP). The general task is to study the solution spaces of source combinatorial\nproblems and to decide whether or not there are sequences of feasible solutions\nthat have special properties. The resulting recongo solver covers all metrics\nof the solver track in the most recent international competition on\ncombinatorial reconfiguration (CoRe Challenge 2022). recongo ranked first in\nthe shortest metric of the single-engine solvers track. In this paper, we\npresent the design and implementation of bounded combinatorial reconfiguration,\nand present an ASP encoding of the independent set reconfiguration problem that\nis one of the most studied combinatorial reconfiguration problems. Finally, we\npresent empirical analysis considering all instances of CoRe Challenge 2022.\n", "date": "2023-07-20", "categories": []} {"id": "2210.12547", "title": "SurCo: Learning Linear Surrogates For Combinatorial Nonlinear\n Optimization Problems", "abstract": " Optimization problems with nonlinear cost functions and combinatorial\nconstraints appear in many real-world applications but remain challenging to\nsolve efficiently compared to their linear counterparts. To bridge this gap, we\npropose $\\textbf{SurCo}$ that learns linear $\\underline{\\text{Sur}}$rogate\ncosts which can be used in existing $\\underline{\\text{Co}}$mbinatorial solvers\nto output good solutions to the original nonlinear combinatorial optimization\nproblem. The surrogate costs are learned end-to-end with nonlinear loss by\ndifferentiating through the linear surrogate solver, combining the flexibility\nof gradient-based methods with the structure of linear combinatorial\noptimization. We propose three $\\texttt{SurCo}$ variants:\n$\\texttt{SurCo}-\\texttt{zero}$ for individual nonlinear problems,\n$\\texttt{SurCo}-\\texttt{prior}$ for problem distributions, and\n$\\texttt{SurCo}-\\texttt{hybrid}$ to combine both distribution and\nproblem-specific information. We give theoretical intuition motivating\n$\\texttt{SurCo}$, and evaluate it empirically. Experiments show that\n$\\texttt{SurCo}$ finds better solutions faster than state-of-the-art and domain\nexpert approaches in real-world optimization problems such as embedding table\nsharding, inverse photonic design, and nonlinear route planning.\n", "date": "2022-10-22", "categories": []} {"id": "2307.09487", "title": "Submodular Maximization under the Intersection of Matroid and Knapsack\n Constraints", "abstract": " Submodular maximization arises in many applications, and has attracted a lot\nof research attentions from various areas such as artificial intelligence,\nfinance and operations research. Previous studies mainly consider only one kind\nof constraint, while many real-world problems often involve several\nconstraints. In this paper, we consider the problem of submodular maximization\nunder the intersection of two commonly used constraints, i.e., $k$-matroid\nconstraint and $m$-knapsack constraint, and propose a new algorithm SPROUT by\nincorporating partial enumeration into the simultaneous greedy framework. We\nprove that SPROUT can achieve a polynomial-time approximation guarantee better\nthan the state-of-the-art algorithms. Then, we introduce the random enumeration\nand smooth techniques into SPROUT to improve its efficiency, resulting in the\nSPROUT++ algorithm, which can keep a similar approximation guarantee.\nExperiments on the applications of movie recommendation and weighted max-cut\ndemonstrate the superiority of SPROUT++ in practice.\n", "date": "2023-07-18", "categories": []} {"id": "2307.10032", "title": "Automatic Conversion of MiniZinc Programs to QUBO", "abstract": " Obtaining Quadratic Unconstrained Binary Optimisation models for various\noptimisation problems, in order to solve those on physical quantum computers\n(such as the the DWave annealers) is nowadays a lengthy and tedious process\nthat requires one to remodel all problem variables as binary variables and\nsqueeze the target function and the constraints into a single quadratic\npolynomial into these new variables.\n We report here on the basis of our automatic converter from MiniZinc to QUBO,\nwhich is able to process a large set of constraint optimisation and constraint\nsatisfaction problems and turn them into equivalent QUBOs, effectively\noptimising the whole process.\n", "date": "2023-07-19", "categories": []} {"id": "2307.09141", "title": "Machine Learning for SAT: Restricted Heuristics and New Graph\n Representations", "abstract": " Boolean satisfiability (SAT) is a fundamental NP-complete problem with many\napplications, including automated planning and scheduling. To solve large\ninstances, SAT solvers have to rely on heuristics, e.g., choosing a branching\nvariable in DPLL and CDCL solvers. Such heuristics can be improved with machine\nlearning (ML) models; they can reduce the number of steps but usually hinder\nthe running time because useful models are relatively large and slow. We\nsuggest the strategy of making a few initial steps with a trained ML model and\nthen releasing control to classical heuristics; this simplifies cold start for\nSAT solving and can decrease both the number of steps and overall runtime, but\nrequires a separate decision of when to release control to the solver.\nMoreover, we introduce a modification of Graph-Q-SAT tailored to SAT problems\nconverted from other domains, e.g., open shop scheduling problems. We validate\nthe feasibility of our approach with random and industrial SAT problems.\n", "date": "2023-07-18", "categories": []} {"id": "2306.01097", "title": "Fast Matrix Multiplication Without Tears: A Constraint Programming\n Approach", "abstract": " It is known that the multiplication of an $N \\times M$ matrix with an $M\n\\times P$ matrix can be performed using fewer multiplications than what the\nnaive $NMP$ approach suggests. The most famous instance of this is Strassen's\nalgorithm for multiplying two $2\\times 2$ matrices in 7 instead of 8\nmultiplications. This gives rise to the constraint satisfaction problem of fast\nmatrix multiplication, where a set of $R < NMP$ multiplication terms must be\nchosen and combined such that they satisfy correctness constraints on the\noutput matrix. Despite its highly combinatorial nature, this problem has not\nbeen exhaustively examined from that perspective, as evidenced for example by\nthe recent deep reinforcement learning approach of AlphaTensor. In this work,\nwe propose a simple yet novel Constraint Programming approach to find\nnon-commutative algorithms for fast matrix multiplication or provide proof of\ninfeasibility otherwise. We propose a set of symmetry-breaking constraints and\nvalid inequalities that are particularly helpful in proving infeasibility. On\nthe feasible side, we find that exploiting solver performance variability in\nconjunction with a sparsity-based problem decomposition enables finding\nsolutions for larger (feasible) instances of fast matrix multiplication. Our\nexperimental results using CP Optimizer demonstrate that we can find fast\nmatrix multiplication algorithms for matrices up to $3\\times 3$ in a short\namount of time.\n", "date": "2023-06-01", "categories": []} {"id": "2307.07322", "title": "A Context-Aware Cutting Plane Selection Algorithm for Mixed-Integer\n Programming", "abstract": " The current cut selection algorithm used in mixed-integer programming solvers\nhas remained largely unchanged since its creation. In this paper, we propose a\nset of new cut scoring measures, cut filtering techniques, and stopping\ncriteria, extending the current state-of-the-art algorithm and obtaining a 5\\%\nperformance improvement for SCIP over the MIPLIB 2017 benchmark set.\n", "date": "2023-07-14", "categories": []} {"id": "2307.06723", "title": "Breaking 3-Factor Approximation for Correlation Clustering in\n Polylogarithmic Rounds", "abstract": " In this paper, we study parallel algorithms for the correlation clustering\nproblem, where every pair of two different entities is labeled with similar or\ndissimilar. The goal is to partition the entities into clusters to minimize the\nnumber of disagreements with the labels. Currently, all efficient parallel\nalgorithms have an approximation ratio of at least 3. In comparison with the\n$1.994+\\epsilon$ ratio achieved by polynomial-time sequential algorithms\n[CLN22], a significant gap exists.\n We propose the first poly-logarithmic depth parallel algorithm that achieves\na better approximation ratio than 3. Specifically, our algorithm computes a\n$(2.4+\\epsilon)$-approximate solution and uses $\\tilde{O}(m^{1.5})$ work.\nAdditionally, it can be translated into a $\\tilde{O}(m^{1.5})$-time sequential\nalgorithm and a poly-logarithmic rounds sublinear-memory MPC algorithm with\n$\\tilde{O}(m^{1.5})$ total memory.\n Our approach is inspired by Awerbuch, Khandekar, and Rao's [AKR12]\nlength-constrained multi-commodity flow algorithm, where we develop an\nefficient parallel algorithm to solve a truncated correlation clustering linear\nprogram of Charikar, Guruswami, and Wirth [CGW05]. Then we show the solution of\nthe truncated linear program can be rounded with a factor of at most 2.4 loss\nby using the framework of [CMSY15]. Such a rounding framework can then be\nimplemented using parallel pivot-based approaches.\n", "date": "2023-07-13", "categories": []} {"id": "2206.04857", "title": "Mixed integer linear optimization formulations for learning optimal\n binary classification trees", "abstract": " Decision trees are powerful tools for classification and regression that\nattract many researchers working in the burgeoning area of machine learning.\nOne advantage of decision trees over other methods is their interpretability,\nwhich is often preferred over other higher accuracy methods that are relatively\nuninterpretable. A binary classification tree has two types of vertices: (i)\nbranching vertices which have exactly two children and where datapoints are\nassessed on a set of discrete features; and (ii) leaf vertices at which\ndatapoints are given a discrete prediction. An optimal binary classification\ntree can be obtained by solving a biobjective optimization problem that seeks\nto (i) maximize the number of correctly classified datapoints and (ii) minimize\nthe number of branching vertices. In this paper, we propose four mixed integer\nlinear optimization (MILO) formulations for designing optimal binary\nclassification trees: two flow-based formulations and two-cut based\nformulations. We provide theoretical comparisons between our proposed\nformulations and the strongest flow-based MILO formulation of Aghaei et al.\n(2021). We conduct experiments on 13 publicly available datasets to show the\nmodels' ability to scale and the strength of a biobjective approach using\nPareto frontiers. Our code and data are available on GitHub.\n", "date": "2022-06-10", "categories": []} {"id": "2212.03712", "title": "Enhanced Multi-Objective A* with Partial Expansion", "abstract": " The Multi-Objective Shortest Path Problem (MO-SPP), typically posed on a\ngraph, determines a set of paths from a start vertex to a destination vertex\nwhile optimizing multiple objectives. In general, there does not exist a single\nsolution path that can simultaneously optimize all the objectives and the\nproblem thus seeks to find a set of so-called Pareto-optimal solutions. To\naddress this problem, several Multi-Objective A* (MOA*) algorithms were\nrecently developed to quickly compute solutions with quality guarantees.\nHowever, these MOA* algorithms often suffer from high memory usage, especially\nwhen the branching factor (i.e. the number of neighbors of any vertex) of the\ngraph is large. This work thus aims at reducing the high memory consumption of\nMOA* with little increase in the runtime. By generalizing and unifying several\nsingle- and multi-objective search algorithms, we develop the Runtime and\nMemory Efficient MOA* (RME-MOA*) approach, which can balance between runtime\nand memory efficiency by tuning two user-defined hyper-parameters.\n", "date": "2022-12-06", "categories": []} {"id": "2305.10292", "title": "Linear Query Approximation Algorithms for Non-monotone Submodular\n Maximization under Knapsack Constraint", "abstract": " This work, for the first time, introduces two constant factor approximation\nalgorithms with linear query complexity for non-monotone submodular\nmaximization over a ground set of size $n$ subject to a knapsack constraint,\n$\\mathsf{DLA}$ and $\\mathsf{RLA}$. $\\mathsf{DLA}$ is a deterministic algorithm\nthat provides an approximation factor of $6+\\epsilon$ while $\\mathsf{RLA}$ is a\nrandomized algorithm with an approximation factor of $4+\\epsilon$. Both run in\n$O(n \\log(1/\\epsilon)/\\epsilon)$ query complexity. The key idea to obtain a\nconstant approximation ratio with linear query lies in: (1) dividing the ground\nset into two appropriate subsets to find the near-optimal solution over these\nsubsets with linear queries, and (2) combining a threshold greedy with\nproperties of two disjoint sets or a random selection process to improve\nsolution quality. In addition to the theoretical analysis, we have evaluated\nour proposed solutions with three applications: Revenue Maximization, Image\nSummarization, and Maximum Weighted Cut, showing that our algorithms not only\nreturn comparative results to state-of-the-art algorithms but also require\nsignificantly fewer queries.\n", "date": "2023-05-17", "categories": []} {"id": "2307.04592", "title": "A Graph Multi-separator Problem for Image Segmentation", "abstract": " We propose a novel abstraction of the image segmentation task in the form of\na combinatorial optimization problem that we call the multi-separator problem.\nFeasible solutions indicate for every pixel whether it belongs to a segment or\na segment separator, and indicate for pairs of pixels whether or not the pixels\nbelong to the same segment. This is in contrast to the closely related lifted\nmulticut problem where every pixel is associated to a segment and no pixel\nexplicitly represents a separating structure. While the multi-separator problem\nis NP-hard, we identify two special cases for which it can be solved\nefficiently. Moreover, we define two local search algorithms for the general\ncase and demonstrate their effectiveness in segmenting simulated volume images\nof foam cells and filaments.\n", "date": "2023-07-10", "categories": []} {"id": "2212.10936", "title": "A Memetic Algorithm with Reinforcement Learning for Sociotechnical\n Production Scheduling", "abstract": " The following interdisciplinary article presents a memetic algorithm with\napplying deep reinforcement learning (DRL) for solving practically oriented\ndual resource constrained flexible job shop scheduling problems (DRC-FJSSP).\nFrom research projects in industry, we recognize the need to consider flexible\nmachines, flexible human workers, worker capabilities, setup and processing\noperations, material arrival times, complex job paths with parallel tasks for\nbill of material (BOM) manufacturing, sequence-dependent setup times and\n(partially) automated tasks in human-machine-collaboration. In recent years,\nthere has been extensive research on metaheuristics and DRL techniques but\nfocused on simple scheduling environments. However, there are few approaches\ncombining metaheuristics and DRL to generate schedules more reliably and\nefficiently. In this paper, we first formulate a DRC-FJSSP to map complex\nindustry requirements beyond traditional job shop models. Then we propose a\nscheduling framework integrating a discrete event simulation (DES) for schedule\nevaluation, considering parallel computing and multicriteria optimization.\nHere, a memetic algorithm is enriched with DRL to improve sequencing and\nassignment decisions. Through numerical experiments with real-world production\ndata, we confirm that the framework generates feasible schedules efficiently\nand reliably for a balanced optimization of makespan (MS) and total tardiness\n(TT). Utilizing DRL instead of random metaheuristic operations leads to better\nresults in fewer algorithm iterations and outperforms traditional approaches in\nsuch complex environments.\n", "date": "2022-12-21", "categories": []} {"id": "2211.14983", "title": "Multiagent Reinforcement Learning for Autonomous Routing and Pickup\n Problem with Adaptation to Variable Demand", "abstract": " We derive a learning framework to generate routing/pickup policies for a\nfleet of autonomous vehicles tasked with servicing stochastically appearing\nrequests on a city map. We focus on policies that 1) give rise to coordination\namongst the vehicles, thereby reducing wait times for servicing requests, 2)\nare non-myopic, and consider a-priori potential future requests, 3) can adapt\nto changes in the underlying demand distribution. Specifically, we are\ninterested in policies that are adaptive to fluctuations of actual demand\nconditions in urban environments, such as on-peak vs. off-peak hours. We\nachieve this through a combination of (i) an online play algorithm that\nimproves the performance of an offline-trained policy, and (ii) an offline\napproximation scheme that allows for adapting to changes in the underlying\ndemand model. In particular, we achieve adaptivity of our learned policy to\ndifferent demand distributions by quantifying a region of validity using the\nq-valid radius of a Wasserstein Ambiguity Set. We propose a mechanism for\nswitching the originally trained offline approximation when the current demand\nis outside the original validity region. In this case, we propose to use an\noffline architecture, trained on a historical demand model that is closer to\nthe current demand in terms of Wasserstein distance. We learn routing and\npickup policies over real taxicab requests in San Francisco with high\nvariability between on-peak and off-peak hours, demonstrating the ability of\nour method to adapt to real fluctuation in demand distributions. Our numerical\nresults demonstrate that our method outperforms alternative rollout-based\nreinforcement learning schemes, as well as other classical methods from\noperations research.\n", "date": "2022-11-28", "categories": []} {"id": "2307.03171", "title": "LEO: Learning Efficient Orderings for Multiobjective Binary Decision\n Diagrams", "abstract": " Approaches based on Binary decision diagrams (BDDs) have recently achieved\nstate-of-the-art results for multiobjective integer programming problems. The\nvariable ordering used in constructing BDDs can have a significant impact on\ntheir size and on the quality of bounds derived from relaxed or restricted BDDs\nfor single-objective optimization problems. We first showcase a similar impact\nof variable ordering on the Pareto frontier (PF) enumeration time for the\nmultiobjective knapsack problem, suggesting the need for deriving variable\nordering methods that improve the scalability of the multiobjective BDD\napproach. To that end, we derive a novel parameter configuration space based on\nvariable scoring functions which are linear in a small set of interpretable and\neasy-to-compute variable features. We show how the configuration space can be\nefficiently explored using black-box optimization, circumventing the curse of\ndimensionality (in the number of variables and objectives), and finding good\norderings that reduce the PF enumeration time. However, black-box optimization\napproaches incur a computational overhead that outweighs the reduction in time\ndue to good variable ordering. To alleviate this issue, we propose LEO, a\nsupervised learning approach for finding efficient variable orderings that\nreduce the enumeration time. Experiments on benchmark sets from the knapsack\nproblem with 3-7 objectives and up to 80 variables show that LEO is ~30-300%\nand ~10-200% faster at PF enumeration than common ordering strategies and\nalgorithm configuration. Our code and instances are available at\nhttps://github.com/khalil-research/leo.\n", "date": "2023-07-06", "categories": []} {"id": "2210.07722", "title": "(1,1)-Cluster Editing is Polynomial-time Solvable", "abstract": " A graph $H$ is a clique graph if $H$ is a vertex-disjoin union of cliques.\nAbu-Khzam (2017) introduced the $(a,d)$-{Cluster Editing} problem, where for\nfixed natural numbers $a,d$, given a graph $G$ and vertex-weights $a^*:\\\nV(G)\\rightarrow \\{0,1,\\dots, a\\}$ and $d^*{}:\\ V(G)\\rightarrow \\{0,1,\\dots,\nd\\}$, we are to decide whether $G$ can be turned into a cluster graph by\ndeleting at most $d^*(v)$ edges incident to every $v\\in V(G)$ and adding at\nmost $a^*(v)$ edges incident to every $v\\in V(G)$. Results by Komusiewicz and\nUhlmann (2012) and Abu-Khzam (2017) provided a dichotomy of complexity (in P or\nNP-complete) of $(a,d)$-{Cluster Editing} for all pairs $a,d$ apart from\n$a=d=1.$ Abu-Khzam (2017) conjectured that $(1,1)$-{Cluster Editing} is in P.\nWe resolve Abu-Khzam's conjecture in affirmative by (i) providing a serious of\nfive polynomial-time reductions to $C_3$-free and $C_4$-free graphs of maximum\ndegree at most 3, and (ii) designing a polynomial-time algorithm for solving\n$(1,1)$-{Cluster Editing} on $C_3$-free and $C_4$-free graphs of maximum degree\nat most 3.\n", "date": "2022-10-14", "categories": []} {"id": "2307.01434", "title": "Learning to Branch in Combinatorial Optimization with Graph Pointer\n Networks", "abstract": " Branch-and-bound is a typical way to solve combinatorial optimization\nproblems. This paper proposes a graph pointer network model for learning the\nvariable selection policy in the branch-and-bound. We extract the graph\nfeatures, global features and historical features to represent the solver\nstate. The proposed model, which combines the graph neural network and the\npointer mechanism, can effectively map from the solver state to the branching\nvariable decisions. The model is trained to imitate the classic strong\nbranching expert rule by a designed top-k Kullback-Leibler divergence loss\nfunction. Experiments on a series of benchmark problems demonstrate that the\nproposed approach significantly outperforms the widely used expert-designed\nbranching rules. Our approach also outperforms the state-of-the-art\nmachine-learning-based branch-and-bound methods in terms of solving speed and\nsearch tree size on all the test instances. In addition, the model can\ngeneralize to unseen instances and scale to larger instances.\n", "date": "2023-07-04", "categories": []} {"id": "2307.00670", "title": "Automatic MILP Solver Configuration By Learning Problem Similarities", "abstract": " A large number of real-world optimization problems can be formulated as Mixed\nInteger Linear Programs (MILP). MILP solvers expose numerous configuration\nparameters to control their internal algorithms. Solutions, and their\nassociated costs or runtimes, are significantly affected by the choice of the\nconfiguration parameters, even when problem instances have the same number of\ndecision variables and constraints. On one hand, using the default solver\nconfiguration leads to suboptimal solutions. On the other hand, searching and\nevaluating a large number of configurations for every problem instance is\ntime-consuming and, in some cases, infeasible. In this study, we aim to predict\nconfiguration parameters for unseen problem instances that yield lower-cost\nsolutions without the time overhead of searching-and-evaluating configurations\nat the solving time. Toward that goal, we first investigate the cost\ncorrelation of MILP problem instances that come from the same distribution when\nsolved using different configurations. We show that instances that have similar\ncosts using one solver configuration also have similar costs using another\nsolver configuration in the same runtime environment. After that, we present a\nmethodology based on Deep Metric Learning to learn MILP similarities that\ncorrelate with their final solutions' costs. At inference time, given a new\nproblem instance, it is first projected into the learned metric space using the\ntrained model, and configuration parameters are instantly predicted using\npreviously-explored configurations from the nearest neighbor instance in the\nlearned embedding space. Empirical results on real-world problem benchmarks\nshow that our method predicts configuration parameters that improve solutions'\ncosts by up to 38% compared to existing approaches.\n", "date": "2023-07-02", "categories": []} {"id": "2307.00783", "title": "Monte Carlo Policy Gradient Method for Binary Optimization", "abstract": " Binary optimization has a wide range of applications in combinatorial\noptimization problems such as MaxCut, MIMO detection, and MaxSAT. However,\nthese problems are typically NP-hard due to the binary constraints. We develop\na novel probabilistic model to sample the binary solution according to a\nparameterized policy distribution. Specifically, minimizing the KL divergence\nbetween the parameterized policy distribution and the Gibbs distributions of\nthe function value leads to a stochastic optimization problem whose policy\ngradient can be derived explicitly similar to reinforcement learning. For\ncoherent exploration in discrete spaces, parallel Markov Chain Monte Carlo\n(MCMC) methods are employed to sample from the policy distribution with\ndiversity and approximate the gradient efficiently. We further develop a filter\nscheme to replace the original objective function by the one with the local\nsearch technique to broaden the horizon of the function landscape. Convergence\nto stationary points in expectation of the policy gradient method is\nestablished based on the concentration inequality for MCMC. Numerical results\nshow that this framework is very promising to provide near-optimal solutions\nfor quite a few binary optimization problems.\n", "date": "2023-07-03", "categories": []} {"id": "2301.12251", "title": "DeciLS-PBO: an Effective Local Search Method for Pseudo-Boolean\n Optimization", "abstract": " Local search is an effective method for solving large-scale combinatorial\noptimization problems, and it has made remarkable progress in recent years\nthrough several subtle mechanisms. In this paper, we found two ways to improve\nthe local search algorithms in solving Pseudo-Boolean Optimization (PBO):\nFirstly, some of those mechanisms such as unit propagation are merely used in\nsolving MaxSAT before, which can be generalized to solve PBO as well; Secondly,\nthe existing local search algorithms utilize the heuristic on variables,\nso-called score, to mainly guide the search. We attempt to gain more insights\ninto the clause, as it plays the role of a middleman who builds a bridge\nbetween variables and the given formula. Hence, we first extended the\ncombination of unit propagation-based decimation algorithm to PBO problem,\ngiving a further generalized definition of unit clause for PBO problem, and\napply it to the existing solver LS-PBO for constructing an initial assignment;\nthen, we introduced a new heuristic on clauses, dubbed care, to set a higher\npriority for the clauses that are less satisfied in current iterations.\nExperiments on benchmarks from the most recent PB Competition, as well as three\nreal-world application benchmarks including minimum-width confidence band,\nwireless sensor network optimization, and seating arrangement problems show\nthat our algorithm DeciLS-PBO has a promising performance compared to the\nstate-of-the-art algorithms.\n", "date": "2023-01-28", "categories": []} {"id": "2304.14712", "title": "Benchmark dataset and instance generator for Real-World\n Three-Dimensional Bin Packing Problems", "abstract": " In this article, a benchmark for real-world bin packing problems is proposed.\nThis dataset consists of 12 instances of varying levels of complexity regarding\nsize (with the number of packages ranging from 38 to 53) and user-defined\nrequirements. In fact, several real-world-oriented restrictions were taken into\naccount to build these instances: i) item and bin dimensions, ii) weight\nrestrictions, iii) affinities among package categories iv) preferences for\npackage ordering and v) load balancing. Besides the data, we also offer an own\ndeveloped Python script for the dataset generation, coined Q4RealBPP-DataGen.\nThe benchmark was initially proposed to evaluate the performance of quantum\nsolvers. Therefore, the characteristics of this set of instances were designed\naccording to the current limitations of quantum devices. Additionally, the\ndataset generator is included to allow the construction of general-purpose\nbenchmarks. The data introduced in this article provides a baseline that will\nencourage quantum computing researchers to work on real-world bin packing\nproblems.\n", "date": "2023-04-28", "categories": []} {"id": "2206.12902", "title": "ETO Meets Scheduling: Learning Key Knowledge from Single-Objective\n Problems to Multi-Objective Problem", "abstract": " Evolutionary transfer optimization(ETO) serves as \"a new frontier in\nevolutionary computation research\", which will avoid zero reuse of experience\nand knowledge from solved problems in traditional evolutionary computation. In\nscheduling applications via ETO, a highly competitive \"meeting\" framework\nbetween them could be constituted towards both intelligent scheduling and green\nscheduling, especially for carbon neutrality within the context of China. To\nthe best of our knowledge, our study on scheduling here, is the 1st work of ETO\nfor complex optimization when multiobjective problem \"meets\" single-objective\nproblems in combinatorial case (not multitasking optimization). More\nspecifically, key knowledge like positional building blocks clustered, could be\nlearned and transferred for permutation flow shop scheduling problem (PFSP).\nEmpirical studies on well-studied benchmarks validate relatively firm\neffectiveness and great potential of our proposed ETO-PFSP framework.\n", "date": "2022-06-26", "categories": []} {"id": "2206.12906", "title": "Towards KAB2S: Learning Key Knowledge from Single-Objective Problems to\n Multi-Objective Problem", "abstract": " As \"a new frontier in evolutionary computation research\", evolutionary\ntransfer optimization(ETO) will overcome the traditional paradigm of zero reuse\nof related experience and knowledge from solved past problems in researches of\nevolutionary computation. In scheduling applications via ETO, a quite appealing\nand highly competitive framework \"meeting\" between them could be formed for\nboth intelligent scheduling and green scheduling, especially for international\npledge of \"carbon neutrality\" from China. To the best of our knowledge, our\npaper on scheduling here, serves as the 1st work of a class of ETO frameworks\nwhen multiobjective optimization problem \"meets\" single-objective optimization\nproblems in discrete case (not multitasking optimization). More specifically,\nkey knowledge conveyed for industrial applications, like positional building\nblocks with genetic algorithm based settings, could be used via the new core\ntransfer mechanism and learning techniques for permutation flow shop scheduling\nproblem(PFSP). Extensive studies on well-studied benchmarks validate firm\neffectiveness and great universality of our proposed ETO-PFSP framework\nempirically. Our investigations (1) enrich the ETO frameworks, (2) contribute\nto the classical and fundamental theory of building block for genetic\nalgorithms and memetic algorithms, and (3) head towards the paradigm shift of\nevolutionary scheduling via learning by proposal and practice of paradigm of\n\"knowledge and building-block based scheduling\" (KAB2S) for \"industrial\nintelligence\" in China.\n", "date": "2022-06-26", "categories": []} {"id": "2306.15182", "title": "Automatic Truss Design with Reinforcement Learning", "abstract": " Truss layout design, namely finding a lightweight truss layout satisfying all\nthe physical constraints, is a fundamental problem in the building industry.\nGenerating the optimal layout is a challenging combinatorial optimization\nproblem, which can be extremely expensive to solve by exhaustive search.\nDirectly applying end-to-end reinforcement learning (RL) methods to truss\nlayout design is infeasible either, since only a tiny portion of the entire\nlayout space is valid under the physical constraints, leading to particularly\nsparse rewards for RL training. In this paper, we develop AutoTruss, a\ntwo-stage framework to efficiently generate both lightweight and valid truss\nlayouts. AutoTruss first adopts Monte Carlo tree search to discover a diverse\ncollection of valid layouts. Then RL is applied to iteratively refine the valid\nsolutions. We conduct experiments and ablation studies in popular truss layout\ndesign test cases in both 2D and 3D settings. AutoTruss outperforms the\nbest-reported layouts by 25.1% in the most challenging 3D test cases, resulting\nin the first effective deep-RL-based approach in the truss layout design\nliterature.\n", "date": "2023-06-27", "categories": []} {"id": "2210.16934", "title": "Learning to Compare Nodes in Branch and Bound with Graph Neural Networks", "abstract": " Branch-and-bound approaches in integer programming require ordering portions\nof the space to explore next, a problem known as node comparison. We propose a\nnew siamese graph neural network model to tackle this problem, where the nodes\nare represented as bipartite graphs with attributes. Similar to prior work, we\ntrain our model to imitate a diving oracle that plunges towards the optimal\nsolution. We evaluate our method by solving the instances in a plain framework\nwhere the nodes are explored according to their rank. On three NP-hard\nbenchmarks chosen to be particularly primal-difficult, our approach leads to\nfaster solving and smaller branch- and-bound trees than the default ranking\nfunction of the open-source solver SCIP, as well as competing machine learning\nmethods. Moreover, these results generalize to instances larger than used for\ntraining. Code for reproducing the experiments can be found at\nhttps://github.com/ds4dm/learn2comparenodes.\n", "date": "2022-10-30", "categories": []} {"id": "2306.06778", "title": "Approximation Algorithms for Fair Range Clustering", "abstract": " This paper studies the fair range clustering problem in which the data points\nare from different demographic groups and the goal is to pick $k$ centers with\nthe minimum clustering cost such that each group is at least minimally\nrepresented in the centers set and no group dominates the centers set. More\nprecisely, given a set of $n$ points in a metric space $(P,d)$ where each point\nbelongs to one of the $\\ell$ different demographics (i.e., $P = P_1 \\uplus P_2\n\\uplus \\cdots \\uplus P_\\ell$) and a set of $\\ell$ intervals $[\\alpha_1,\n\\beta_1], \\cdots, [\\alpha_\\ell, \\beta_\\ell]$ on desired number of centers from\neach group, the goal is to pick a set of $k$ centers $C$ with minimum\n$\\ell_p$-clustering cost (i.e., $(\\sum_{v\\in P} d(v,C)^p)^{1/p}$) such that for\neach group $i\\in \\ell$, $|C\\cap P_i| \\in [\\alpha_i, \\beta_i]$. In particular,\nthe fair range $\\ell_p$-clustering captures fair range $k$-center, $k$-median\nand $k$-means as its special cases. In this work, we provide efficient constant\nfactor approximation algorithms for fair range $\\ell_p$-clustering for all\nvalues of $p\\in [1,\\infty)$.\n", "date": "2023-06-11", "categories": []} {"id": "2305.05119", "title": "Flexible Job Shop Scheduling via Dual Attention Network Based\n Reinforcement Learning", "abstract": " Flexible manufacturing has given rise to complex scheduling problems such as\nthe flexible job shop scheduling problem (FJSP). In FJSP, operations can be\nprocessed on multiple machines, leading to intricate relationships between\noperations and machines. Recent works have employed deep reinforcement learning\n(DRL) to learn priority dispatching rules (PDRs) for solving FJSP. However, the\nquality of solutions still has room for improvement relative to that by the\nexact methods such as OR-Tools. To address this issue, this paper presents a\nnovel end-to-end learning framework that weds the merits of self-attention\nmodels for deep feature extraction and DRL for scalable decision-making. The\ncomplex relationships between operations and machines are represented precisely\nand concisely, for which a dual-attention network (DAN) comprising several\ninterconnected operation message attention blocks and machine message attention\nblocks is proposed. The DAN exploits the complicated relationships to construct\nproduction-adaptive operation and machine features to support high-quality\ndecisionmaking. Experimental results using synthetic data as well as public\nbenchmarks corroborate that the proposed approach outperforms both traditional\nPDRs and the state-of-the-art DRL method. Moreover, it achieves results\ncomparable to exact methods in certain cases and demonstrates favorable\ngeneralization ability to large-scale and real-world unseen FJSP tasks.\n", "date": "2023-05-09", "categories": []} {"id": "2305.19587", "title": "Towards Omni-generalizable Neural Methods for Vehicle Routing Problems", "abstract": " Learning heuristics for vehicle routing problems (VRPs) has gained much\nattention due to the less reliance on hand-crafted rules. However, existing\nmethods are typically trained and tested on the same task with a fixed size and\ndistribution (of nodes), and hence suffer from limited generalization\nperformance. This paper studies a challenging yet realistic setting, which\nconsiders generalization across both size and distribution in VRPs. We propose\na generic meta-learning framework, which enables effective training of an\ninitialized model with the capability of fast adaptation to new tasks during\ninference. We further develop a simple yet efficient approximation method to\nreduce the training overhead. Extensive experiments on both synthetic and\nbenchmark instances of the traveling salesman problem (TSP) and capacitated\nvehicle routing problem (CVRP) demonstrate the effectiveness of our method. The\ncode is available at: https://github.com/RoyalSkye/Omni-VRP.\n", "date": "2023-05-31", "categories": []} {"id": "2211.03074", "title": "A Survey on Influence Maximization: From an ML-Based Combinatorial\n Optimization", "abstract": " Influence Maximization (IM) is a classical combinatorial optimization\nproblem, which can be widely used in mobile networks, social computing, and\nrecommendation systems. It aims at selecting a small number of users such that\nmaximizing the influence spread across the online social network. Because of\nits potential commercial and academic value, there are a lot of researchers\nfocusing on studying the IM problem from different perspectives. The main\nchallenge comes from the NP-hardness of the IM problem and \\#P-hardness of\nestimating the influence spread, thus traditional algorithms for overcoming\nthem can be categorized into two classes: heuristic algorithms and\napproximation algorithms. However, there is no theoretical guarantee for\nheuristic algorithms, and the theoretical design is close to the limit.\nTherefore, it is almost impossible to further optimize and improve their\nperformance. With the rapid development of artificial intelligence, the\ntechnology based on Machine Learning (ML) has achieved remarkable achievements\nin many fields. In view of this, in recent years, a number of new methods have\nemerged to solve combinatorial optimization problems by using ML-based\ntechniques. These methods have the advantages of fast solving speed and strong\ngeneralization ability to unknown graphs, which provide a brand-new direction\nfor solving combinatorial optimization problems. Therefore, we abandon the\ntraditional algorithms based on iterative search and review the recent\ndevelopment of ML-based methods, especially Deep Reinforcement Learning, to\nsolve the IM problem and other variants in social networks. We focus on\nsummarizing the relevant background knowledge, basic principles, common\nmethods, and applied research. Finally, the challenges that need to be solved\nurgently in future IM research are pointed out.\n", "date": "2022-11-06", "categories": []} {"id": "2210.12453", "title": "NeuroPrim: An Attention-based Model for Solving NP-hard Spanning Tree\n Problems", "abstract": " Spanning tree problems with specialized constraints can be difficult to solve\nin real-world scenarios, often requiring intricate algorithmic design and\nexponential time. Recently, there has been growing interest in end-to-end deep\nneural networks for solving routing problems. However, such methods typically\nproduce sequences of vertices, which makes it difficult to apply them to\ngeneral combinatorial optimization problems where the solution set consists of\nedges, as in various spanning tree problems. In this paper, we propose\nNeuroPrim, a novel framework for solving various spanning tree problems by\ndefining a Markov Decision Process (MDP) for general combinatorial optimization\nproblems on graphs. Our approach reduces the action and state space using\nPrim's algorithm and trains the resulting model using REINFORCE. We apply our\nframework to three difficult problems on Euclidean space: the\nDegree-constrained Minimum Spanning Tree (DCMST) problem, the Minimum Routing\nCost Spanning Tree (MRCST) problem, and the Steiner Tree Problem in graphs\n(STP). Experimental results on literature instances demonstrate that our model\noutperforms strong heuristics and achieves small optimality gaps of up to 250\nvertices. Additionally, we find that our model has strong generalization\nability, with no significant degradation observed on problem instances as large\nas 1000. Our results suggest that our framework can be effective for solving a\nwide range of combinatorial optimization problems beyond spanning tree\nproblems.\n", "date": "2022-10-22", "categories": []} {"id": "2306.05747", "title": "An End-to-End Reinforcement Learning Approach for Job-Shop Scheduling\n Problems Based on Constraint Programming", "abstract": " Constraint Programming (CP) is a declarative programming paradigm that allows\nfor modeling and solving combinatorial optimization problems, such as the\nJob-Shop Scheduling Problem (JSSP). While CP solvers manage to find optimal or\nnear-optimal solutions for small instances, they do not scale well to large\nones, i.e., they require long computation times or yield low-quality solutions.\nTherefore, real-world scheduling applications often resort to fast,\nhandcrafted, priority-based dispatching heuristics to find a good initial\nsolution and then refine it using optimization methods.\n This paper proposes a novel end-to-end approach to solving scheduling\nproblems by means of CP and Reinforcement Learning (RL). In contrast to\nprevious RL methods, tailored for a given problem by including procedural\nsimulation algorithms, complex feature engineering, or handcrafted reward\nfunctions, our neural-network architecture and training algorithm merely\nrequire a generic CP encoding of some scheduling problem along with a set of\nsmall instances. Our approach leverages existing CP solvers to train an agent\nlearning a Priority Dispatching Rule (PDR) that generalizes well to large\ninstances, even from separate datasets. We evaluate our method on seven JSSP\ndatasets from the literature, showing its ability to find higher-quality\nsolutions for very large instances than obtained by static PDRs and by a CP\nsolver within the same time limit.\n", "date": "2023-06-09", "categories": []} {"id": "2306.05865", "title": "Faster Discrete Convex Function Minimization with Predictions: The\n M-Convex Case", "abstract": " Recent years have seen a growing interest in accelerating optimization\nalgorithms with machine-learned predictions. Sakaue and Oki (NeurIPS 2022) have\ndeveloped a general framework that warm-starts the L-convex function\nminimization method with predictions, revealing the idea's usefulness for\nvarious discrete optimization problems. In this paper, we present a framework\nfor using predictions to accelerate M-convex function minimization, thus\ncomplementing previous research and extending the range of discrete\noptimization algorithms that can benefit from predictions. Our framework is\nparticularly effective for an important subclass called laminar convex\nminimization, which appears in many operations research applications. Our\nmethods can improve time complexity bounds upon the best worst-case results by\nusing predictions and even have potential to go beyond a lower-bound result.\n", "date": "2023-06-09", "categories": []} {"id": "2306.02688", "title": "Meta-SAGE: Scale Meta-Learning Scheduled Adaptation with Guided\n Exploration for Mitigating Scale Shift on Combinatorial Optimization", "abstract": " This paper proposes Meta-SAGE, a novel approach for improving the scalability\nof deep reinforcement learning models for combinatorial optimization (CO)\ntasks. Our method adapts pre-trained models to larger-scale problems in test\ntime by suggesting two components: a scale meta-learner (SML) and scheduled\nadaptation with guided exploration (SAGE). First, SML transforms the context\nembedding for subsequent adaptation of SAGE based on scale information. Then,\nSAGE adjusts the model parameters dedicated to the context embedding for a\nspecific instance. SAGE introduces locality bias, which encourages selecting\nnearby locations to determine the next location. The locality bias gradually\ndecays as the model is adapted to the target instance. Results show that\nMeta-SAGE outperforms previous adaptation methods and significantly improves\nscalability in representative CO tasks. Our source code is available at\nhttps://github.com/kaist-silab/meta-sage\n", "date": "2023-06-05", "categories": []} {"id": "2306.04403", "title": "Policy-Based Self-Competition for Planning Problems", "abstract": " AlphaZero-type algorithms may stop improving on single-player tasks in case\nthe value network guiding the tree search is unable to approximate the outcome\nof an episode sufficiently well. One technique to address this problem is\ntransforming the single-player task through self-competition. The main idea is\nto compute a scalar baseline from the agent's historical performances and to\nreshape an episode's reward into a binary output, indicating whether the\nbaseline has been exceeded or not. However, this baseline only carries limited\ninformation for the agent about strategies how to improve. We leverage the idea\nof self-competition and directly incorporate a historical policy into the\nplanning process instead of its scalar performance. Based on the recently\nintroduced Gumbel AlphaZero (GAZ), we propose our algorithm GAZ 'Play-to-Plan'\n(GAZ PTP), in which the agent learns to find strong trajectories by planning\nagainst possible strategies of its past self. We show the effectiveness of our\napproach in two well-known combinatorial optimization problems, the Traveling\nSalesman Problem and the Job-Shop Scheduling Problem. With only half of the\nsimulation budget for search, GAZ PTP consistently outperforms all selected\nsingle-player variants of GAZ.\n", "date": "2023-06-07", "categories": []} {"id": "2212.03178", "title": "Longest Common Substring in Longest Common Subsequence's Solution\n Service: A Novel Hyper-Heuristic", "abstract": " The Longest Common Subsequence (LCS) is the problem of finding a subsequence\namong a set of strings that has two properties of being common to all and is\nthe longest. The LCS has applications in computational biology and text\nediting, among many others. Due to the NP-hardness of the general longest\ncommon subsequence, numerous heuristic algorithms and solvers have been\nproposed to give the best possible solution for different sets of strings. None\nof them has the best performance for all types of sets. In addition, there is\nno method to specify the type of a given set of strings. Besides that, the\navailable hyper-heuristic is not efficient and fast enough to solve this\nproblem in real-world applications. This paper proposes a novel hyper-heuristic\nto solve the longest common subsequence problem using a novel criterion to\nclassify a set of strings based on their similarity. To do this, we offer a\ngeneral stochastic framework to identify the type of a given set of strings.\nFollowing that, we introduce the set similarity dichotomizer ($S^2D$) algorithm\nbased on the framework that divides the type of sets into two. This algorithm\nis introduced for the first time in this paper and opens a new way to go beyond\nthe current LCS solvers. Then, we present a novel hyper-heuristic that exploits\nthe $S^2D$ and one of the internal properties of the set to choose the best\nmatching heuristic among a set of heuristics. We compare the results on\nbenchmark datasets with the best heuristics and hyper-heuristics. The results\nshow a higher performance of our proposed hyper-heuristic in both quality of\nsolutions and run time factors.\n", "date": "2022-12-03", "categories": []} {"id": "2302.14471", "title": "Safe Peeling for L0-Regularized Least-Squares with supplementary\n material", "abstract": " We introduce a new methodology dubbed ``safe peeling'' to accelerate the\nresolution of L0-regularized least-squares problems via a Branch-and-Bound\n(BnB) algorithm. Our procedure enables to tighten the convex relaxation\nconsidered at each node of the BnB decision tree and therefore potentially\nallows for more aggressive pruning. Numerical simulations show that our\nproposed methodology leads to significant gains in terms of number of nodes\nexplored and overall solving time.s show that our proposed methodology leads to\nsignificant gains in terms of number of nodes explored and overall solving\ntime.\n", "date": "2023-02-28", "categories": []} {"id": "2306.01021", "title": "A Virtual-Force Based Swarm Algorithm for Balanced Circular Bin Packing\n Problems", "abstract": " Balanced circular bin packing problems consist in positioning a given number\nof weighted circles in order to minimize the radius of a circular container\nwhile satisfying equilibrium constraints. These problems are NP-hard, highly\nconstrained and dimensional. This paper describes a swarm algorithm based on a\nvirtual-force system in order to solve balanced circular bin packing problems.\nIn the proposed approach, a system of forces is applied to each component\nallowing to take into account the constraints and minimizing the objective\nfunction using the fundamental principle of dynamics. The proposed algorithm is\nexperimented and validated on benchmarks of various balanced circular bin\npacking problems with up to 300 circles. The reported results allow to assess\nthe effectiveness of the proposed approach compared to existing results from\nthe literature.\n", "date": "2023-06-01", "categories": []} {"id": "2306.03409", "title": "Rigorous Runtime Analysis of MOEA/D for Solving Multi-Objective Minimum\n Weight Base Problems", "abstract": " We study the multi-objective minimum weight base problem, an abstraction of\nclassical NP-hard combinatorial problems such as the multi-objective minimum\nspanning tree problem. We prove some important properties of the convex hull of\nthe non-dominated front, such as its approximation quality and an upper bound\non the number of extreme points. Using these properties, we give the first\nrun-time analysis of the MOEA/D algorithm for this problem, an evolutionary\nalgorithm that effectively optimizes by decomposing the objectives into\nsingle-objective components. We show that the MOEA/D, given an appropriate\ndecomposition setting, finds all extreme points within expected fixed-parameter\npolynomial time in the oracle model, the parameter being the number of\nobjectives. Experiments are conducted on random bi-objective minimum spanning\ntree instances, and the results agree with our theoretical findings.\nFurthermore, compared with a previously studied evolutionary algorithm for the\nproblem GSEMO, MOEA/D finds all extreme points much faster across all\ninstances.\n", "date": "2023-06-06", "categories": []} {"id": "2306.03434", "title": "Learning-Based Heuristic for Combinatorial Optimization of the Minimum\n Dominating Set Problem using Graph Convolutional Networks", "abstract": " A dominating set of a graph $\\mathcal{G=(V, E)}$ is a subset of vertices\n$S\\subseteq\\mathcal{V}$ such that every vertex $v\\in \\mathcal{V} \\setminus S$\noutside the dominating set is adjacent to a vertex $u\\in S$ within the set. The\nminimum dominating set problem seeks to find a dominating set of minimum\ncardinality and is a well-established NP-hard combinatorial optimization\nproblem. We propose a novel learning-based heuristic approach to compute\nsolutions for the minimum dominating set problem using graph convolutional\nnetworks. We conduct an extensive experimental evaluation of the proposed\nmethod on a combination of randomly generated graphs and real-world graph\ndatasets. Our results indicate that the proposed learning-based approach can\noutperform a classical greedy approximation algorithm. Furthermore, we\ndemonstrate the generalization capability of the graph convolutional network\nacross datasets and its ability to scale to graphs of higher order than those\non which it was trained. Finally, we utilize the proposed learning-based\nheuristic in an iterative greedy algorithm, achieving state-of-the-art\nperformance in the computation of dominating sets.\n", "date": "2023-06-06", "categories": []} {"id": "2211.15368", "title": "Arbitrarily Large Labelled Random Satisfiability Formulas for Machine\n Learning Training", "abstract": " Applying deep learning to solve real-life instances of hard combinatorial\nproblems has tremendous potential. Research in this direction has focused on\nthe Boolean satisfiability (SAT) problem, both because of its theoretical\ncentrality and practical importance. A major roadblock faced, though, is that\ntraining sets are restricted to random formulas of size several orders of\nmagnitude smaller than formulas of practical interest, raising serious concerns\nabout generalization. This is because labeling random formulas of increasing\nsize rapidly becomes intractable. By exploiting the probabilistic method in a\nfundamental way, we remove this roadblock entirely: we show how to generate\ncorrectly labeled random formulas of any desired size, without having to solve\nthe underlying decision problem. Moreover, the difficulty of the classification\ntask for the formulas produced by our generator is tunable by varying a simple\nscalar parameter. This opens up an entirely new level of sophistication for the\nmachine learning methods that can be brought to bear on Satisfiability. Using\nour generator, we train existing state-of-the-art models for the task of\npredicting satisfiability on formulas with 10,000 variables. We find that they\ndo no better than random guessing. As a first indication of what can be\nachieved with the new generator, we present a novel classifier that performs\nsignificantly better than random guessing 99% on the same datasets, for most\ndifficulty levels. Crucially, unlike past approaches that learn based on\nsyntactic features of a formula, our classifier performs its learning on a\nshort prefix of a solver's computation, an approach that we expect to be of\nindependent interest.\n", "date": "2022-11-21", "categories": []} {"id": "2301.12159", "title": "ClusterFuG: Clustering Fully connected Graphs by Multicut", "abstract": " We propose a graph clustering formulation based on multicut (a.k.a. weighted\ncorrelation clustering) on the complete graph. Our formulation does not need\nspecification of the graph topology as in the original sparse formulation of\nmulticut, making our approach simpler and potentially better performing. In\ncontrast to unweighted correlation clustering we allow for a more expressive\nweighted cost structure. In dense multicut, the clustering objective is given\nin a factorized form as inner products of node feature vectors. This allows for\nan efficient formulation and inference in contrast to multicut/weighted\ncorrelation clustering, which has at least quadratic representation and\ncomputation complexity when working on the complete graph. We show how to\nrewrite classical greedy algorithms for multicut in our dense setting and how\nto modify them for greater efficiency and solution quality. In particular, our\nalgorithms scale to graphs with tens of thousands of nodes. Empirical evidence\non instance segmentation on Cityscapes and clustering of ImageNet datasets\nshows the merits of our approach.\n", "date": "2023-01-28", "categories": []} {"id": "2305.06167", "title": "K-SpecPart: Supervised embedding algorithms and cut overlay for improved\n hypergraph partitioning", "abstract": " State-of-the-art hypergraph partitioners follow the multilevel paradigm that\nconstructs multiple levels of progressively coarser hypergraphs that are used\nto drive cut refinement on each level of the hierarchy. Multilevel partitioners\nare subject to two limitations: (i) hypergraph coarsening processes rely on\nlocal neighborhood structure without fully considering the global structure of\nthe hypergraph; and (ii) refinement heuristics risk entrapment in local minima.\nIn this paper, we describe K-SpecPart, a supervised spectral framework for\nmulti-way partitioning that directly tackles these two limitations. K-SpecPart\nrelies on the computation of generalized eigenvectors and supervised\ndimensionality reduction techniques to generate vertex embeddings. These are\ncomputational primitives that are fast and capture global structural properties\nof the hypergraph that are not explicitly considered by existing partitioners.\nK-SpecPart then converts the vertex embeddings into multiple partitioning\nsolutions. K-SpecPart introduces the idea of ''ensembling'' multiple solutions\nvia a cut-overlay clustering technique that often enables the use of\ncomputationally demanding partitioning methods such as ILP (integer linear\nprogramming). Using the output of a standard partitioner as a supervision hint,\nK-SpecPart effectively combines the strengths of established multilevel\npartitioning techniques with the benefits of spectral graph theory and other\ncombinatorial algorithms. K-SpecPart significantly extends ideas and algorithms\nthat first appeared in our previous work on the bipartitioner SpecPart. Our\nexperiments demonstrate the effectiveness of K-SpecPart. For bipartitioning,\nK-SpecPart produces solutions with up to 15% cutsize improvement over SpecPart.\nFor multi-way partitioning, K-SpecPart produces solutions with up to 20%\ncutsize improvement over leading partitioners hMETIS and KaHyPar.\n", "date": "2023-05-07", "categories": []} {"id": "2305.13823", "title": "XRoute Environment: A Novel Reinforcement Learning Environment for\n Routing", "abstract": " Routing is a crucial and time-consuming stage in modern design automation\nflow for advanced technology nodes. Great progress in the field of\nreinforcement learning makes it possible to use those approaches to improve the\nrouting quality and efficiency. However, the scale of the routing problems\nsolved by reinforcement learning-based methods in recent studies is too small\nfor these methods to be used in commercial EDA tools. We introduce the XRoute\nEnvironment, a new reinforcement learning environment where agents are trained\nto select and route nets in an advanced, end-to-end routing framework. Novel\nalgorithms and ideas can be quickly tested in a safe and reproducible manner in\nit. The resulting environment is challenging, easy to use, customize and add\nadditional scenarios, and it is available under a permissive open-source\nlicense. In addition, it provides support for distributed deployment and\nmulti-instance experiments. We propose two tasks for learning and build a\nfull-chip test bed with routing benchmarks of various region sizes. We also\npre-define several static routing regions with different pin density and number\nof nets for easier learning and testing. For net ordering task, we report\nbaseline results for two widely used reinforcement learning algorithms (PPO and\nDQN) and one searching-based algorithm (TritonRoute). The XRoute Environment\nwill be available at https://github.com/xplanlab/xroute_env.\n", "date": "2023-05-23", "categories": []} {"id": "2306.02296", "title": "Onsite Job Scheduling by Adaptive Genetic Algorithm", "abstract": " Onsite Job Scheduling is a specialized variant of Vehicle Routing Problem\n(VRP) with multiple depots. The objective of this problem is to execute jobs\nrequested by customers, belonging to different geographic locations by a\nlimited number of technicians, with minimum travel and overtime of technicians.\nEach job is expected to be completed within a specified time limit according to\nthe service level agreement with customers. Each technician is assumed to start\nfrom a base location, serve several customers and return to the starting place.\nTechnicians are allotted jobs based on their skill sets, expertise levels of\neach skill and availability slots. Although there are considerable number of\nliteratures on VRP we do not see any explicit work related to Onsite Job\nScheduling. In this paper we have proposed an Adaptive Genetic Algorithm to\nsolve the scheduling problem. We found an optimized travel route for a\nsubstantial number of jobs and technicians, minimizing travel distance,\novertime duration as well as meeting constraints related to SLA.\n", "date": "2023-06-04", "categories": []} {"id": "2212.02094", "title": "Learning Physically Realizable Skills for Online Packing of General 3D\n Shapes", "abstract": " We study the problem of learning online packing skills for irregular 3D\nshapes, which is arguably the most challenging setting of bin packing problems.\nThe goal is to consecutively move a sequence of 3D objects with arbitrary\nshapes into a designated container with only partial observations of the object\nsequence. Meanwhile, we take physical realizability into account, involving\nphysics dynamics and constraints of a placement. The packing policy should\nunderstand the 3D geometry of the object to be packed and make effective\ndecisions to accommodate it in the container in a physically realizable way. We\npropose a Reinforcement Learning (RL) pipeline to learn the policy. The complex\nirregular geometry and imperfect object placement together lead to huge\nsolution space. Direct training in such space is prohibitively data intensive.\nWe instead propose a theoretically-provable method for candidate action\ngeneration to reduce the action space of RL and the learning burden. A\nparameterized policy is then learned to select the best placement from the\ncandidates. Equipped with an efficient method of asynchronous RL acceleration\nand a data preparation process of simulation-ready training sequences, a mature\npacking policy can be trained in a physics-based environment within 48 hours.\nThrough extensive evaluation on a variety of real-life shape datasets and\ncomparisons with state-of-the-art baselines, we demonstrate that our method\noutperforms the best-performing baseline on all datasets by at least 12.8% in\nterms of packing utility.\n", "date": "2022-12-05", "categories": []} {"id": "2212.08183", "title": "Local Branching Relaxation Heuristics for Integer Linear Programs", "abstract": " Large Neighborhood Search (LNS) is a popular heuristic algorithm for solving\ncombinatorial optimization problems (COP). It starts with an initial solution\nto the problem and iteratively improves it by searching a large neighborhood\naround the current best solution. LNS relies on heuristics to select\nneighborhoods to search in. In this paper, we focus on designing effective and\nefficient heuristics in LNS for integer linear programs (ILP) since a wide\nrange of COPs can be represented as ILPs. Local Branching (LB) is a heuristic\nthat selects the neighborhood that leads to the largest improvement over the\ncurrent solution in each iteration of LNS. LB is often slow since it needs to\nsolve an ILP of the same size as input. Our proposed heuristics, LB-RELAX and\nits variants, use the linear programming relaxation of LB to select\nneighborhoods. Empirically, LB-RELAX and its variants compute as effective\nneighborhoods as LB but run faster. They achieve state-of-the-art anytime\nperformance on several ILP benchmarks.\n", "date": "2022-12-15", "categories": []} {"id": "2306.00720", "title": "Neural Bee Colony Optimization: A Case Study in Public Transit Network\n Design", "abstract": " In this work we explore the combination of metaheuristics and learned neural\nnetwork solvers for combinatorial optimization. We do this in the context of\nthe transit network design problem, a uniquely challenging combinatorial\noptimization problem with real-world importance. We train a neural network\npolicy to perform single-shot planning of individual transit routes, and then\nincorporate it as one of several sub-heuristics in a modified Bee Colony\nOptimization (BCO) metaheuristic algorithm. Our experimental results\ndemonstrate that this hybrid algorithm outperforms the learned policy alone by\nup to 20% and the original BCO algorithm by up to 53% on realistic problem\ninstances. We perform a set of ablations to study the impact of each component\nof the modified algorithm.\n", "date": "2023-05-18", "categories": []} {"id": "2305.18859", "title": "Large-scale Ridesharing DARP Instances Based on Real Travel Demand", "abstract": " Accurately predicting the real-life performance of algorithms solving the\nDial-a-Ride Problem (DARP) in the context of Mobility on Demand (MoD) systems\nwith ridesharing requires evaluating them on representative instances. However,\nthe benchmarking of state-of-the-art DARP solution methods has been limited to\nsmall, artificial instances or outdated non-public instances, hindering direct\ncomparisons. With the rise of large MoD systems and the availability of open\ntravel demand datasets for many US cities, there is now an opportunity to\nevaluate these algorithms on standardized, realistic, and representative\ninstances. Despite the significant challenges involved in processing obfuscated\nand diverse datasets, we have developed a methodology using which we have\ncreated a comprehensive set of large-scale demand instances based on real-world\ndata. These instances cover diverse use cases, one of which is demonstrated in\nan evaluation of two established DARP methods: the insertion heuristic and\noptimal vehicle-group assignment method. We publish the full results of both\nmethods in a standardized format. The results show significant differences\nbetween areas in all measured quantities, emphasizing the importance of\nevaluating methods across different cities.\n", "date": "2023-05-30", "categories": []} {"id": "2305.19077", "title": "DHRL-FNMR: An Intelligent Multicast Routing Approach Based on Deep\n Hierarchical Reinforcement Learning in SDN", "abstract": " The optimal multicast tree problem in the Software-Defined Networking (SDN)\nmulticast routing is an NP-hard combinatorial optimization problem. Although\nexisting SDN intelligent solution methods, which are based on deep\nreinforcement learning, can dynamically adapt to complex network link state\nchanges, these methods are plagued by problems such as redundant branches,\nlarge action space, and slow agent convergence. In this paper, an SDN\nintelligent multicast routing algorithm based on deep hierarchical\nreinforcement learning is proposed to circumvent the aforementioned problems.\nFirst, the multicast tree construction problem is decomposed into two\nsub-problems: the fork node selection problem and the construction of the\noptimal path from the fork node to the destination node. Second, based on the\ninformation characteristics of SDN global network perception, the multicast\ntree state matrix, link bandwidth matrix, link delay matrix, link packet loss\nrate matrix, and sub-goal matrix are designed as the state space of intrinsic\nand meta controllers. Then, in order to mitigate the excessive action space,\nour approach constructs different action spaces at the upper and lower levels.\nThe meta-controller generates an action space using network nodes to select the\nfork node, and the intrinsic controller uses the adjacent edges of the current\nnode as its action space, thus implementing four different action selection\nstrategies in the construction of the multicast tree. To facilitate the\nintelligent agent in constructing the optimal multicast tree with greater\nspeed, we developed alternative reward strategies that distinguish between\nsingle-step node actions and multi-step actions towards multiple destination\nnodes.\n", "date": "2023-05-30", "categories": []} {"id": "2305.16903", "title": "Submodular Minimax Optimization: Finding Effective Sets", "abstract": " Despite the rich existing literature about minimax optimization in continuous\nsettings, only very partial results of this kind have been obtained for\ncombinatorial settings. In this paper, we fill this gap by providing a\ncharacterization of submodular minimax optimization, the problem of finding a\nset (for either the min or the max player) that is effective against every\npossible response. We show when and under what conditions we can find such\nsets. We also demonstrate how minimax submodular optimization provides robust\nsolutions for downstream machine learning applications such as (i) efficient\nprompt engineering for question answering, (ii) prompt engineering for dialog\nstate tracking, (iii) identifying robust waiting locations for ride-sharing,\n(iv) ride-share difficulty kernelization, and (v) finding adversarial images.\nOur experiments demonstrate that our proposed algorithms consistently\noutperform other baselines.\n", "date": "2023-05-26", "categories": []} {"id": "2206.06789", "title": "Grid-SiPhyR: An end-to-end learning to optimize framework for\n combinatorial problems in power systems", "abstract": " Mixed integer problems are ubiquitous in decision making, from discrete\ndevice settings and design parameters, unit production, and on/off or yes/no\ndecision in switches, routing, and social networks. Despite their prevalence,\nclassical optimization approaches for combinatorial optimization remain\nprohibitively slow for fast and accurate decision making in dynamic and\nsafety-critical environments with hard constraints. To address this gap, we\npropose SiPhyR (pronounced: cipher), a physics-informed machine learning\nframework for end-to-end learning to optimize for combinatorial problems.\nSiPhyR employs a novel physics-informed rounding approach to tackle the\nchallenge of combinatorial optimization within a differentiable framework that\nhas certified satisfiability of safety-critical constraints. We demonstrate the\neffectiveness of SiPhyR on an emerging paradigm for clean energy systems:\ndynamic reconfiguration, where the topology of the electric grid and power flow\nare optimized so as to maintain a safe and reliable power grid in the presence\nof intermittent renewable generation. Offline training of the unsupervised\nframework on representative load and generation data makes dynamic decision\nmaking via the online application of Grid-SiPhyR computationally feasible.\n", "date": "2022-06-11", "categories": []} {"id": "2305.16013", "title": "Online and Streaming Algorithms for Constrained $k$-Submodular\n Maximization", "abstract": " Constrained $k$-submodular maximization is a general framework that captures\nmany discrete optimization problems such as ad allocation, influence\nmaximization, personalized recommendation, and many others. In many of these\napplications, datasets are large or decisions need to be made in an online\nmanner, which motivates the development of efficient streaming and online\nalgorithms. In this work, we develop single-pass streaming and online\nalgorithms for constrained $k$-submodular maximization with both monotone and\ngeneral (possibly non-monotone) objectives subject to cardinality and knapsack\nconstraints. Our algorithms achieve provable constant-factor approximation\nguarantees which improve upon the state of the art in almost all settings.\nMoreover, they are combinatorial and very efficient, and have optimal space and\nrunning time. We experimentally evaluate our algorithms on instances for ad\nallocation and other applications, where we observe that our algorithms are\nefficient and scalable, and construct solutions that are comparable in value to\noffline greedy algorithms.\n", "date": "2023-05-25", "categories": []} {"id": "2305.16191", "title": "UpMax: User partitioning for MaxSAT", "abstract": " It has been shown that Maximum Satisfiability (MaxSAT) problem instances can\nbe effectively solved by partitioning the set of soft clauses into several\ndisjoint sets. The partitioning methods can be based on clause weights (e.g.,\nstratification) or based on graph representations of the formula. Afterwards, a\nmerge procedure is applied to guarantee that an optimal solution is found.\n This paper proposes a new framework called UpMax that decouples the\npartitioning procedure from the MaxSAT solving algorithms. As a result, new\npartitioning procedures can be defined independently of the MaxSAT algorithm to\nbe used. Moreover, this decoupling also allows users that build new MaxSAT\nformulas to propose partition schemes based on knowledge of the problem to be\nsolved. We illustrate this approach using several problems and show that\npartitioning has a large impact on the performance of unsatisfiability-based\nMaxSAT algorithms.\n", "date": "2023-05-25", "categories": []} {"id": "2305.16303", "title": "Fine-Grained Complexity Analysis of Multi-Agent Path Finding on 2D Grids", "abstract": " Multi-Agent Path Finding (MAPF) is a fundamental motion coordination problem\narising in multi-agent systems with a wide range of applications. The problem's\nintractability has led to extensive research on improving the scalability of\nsolvers for it. Since optimal solvers can struggle to scale, a major challenge\nthat arises is understanding what makes MAPF hard. We tackle this challenge\nthrough a fine-grained complexity analysis of time-optimal MAPF on 2D grids,\nthereby closing two gaps and identifying a new tractability frontier. First, we\nshow that 2-colored MAPF, i.e., where the agents are divided into two teams,\neach with its own set of targets, remains NP-hard. Second, for the flowtime\nobjective (also called sum-of-costs), we show that it remains NP-hard to find a\nsolution in which agents have an individually optimal cost, which we call an\nindividually optimal solution. The previously tightest results for these MAPF\nvariants are for (non-grid) planar graphs. We use a single hardness\nconstruction that replaces, strengthens, and unifies previous proofs. We\nbelieve that it is also simpler than previous proofs for the planar case as it\nemploys minimal gadgets that enable its full visualization in one figure.\nFinally, for the flowtime objective, we establish a tractability frontier based\non the number of directions agents can move in. Namely, we complement our\nhardness result, which holds for three directions, with an efficient algorithm\nfor finding an individually optimal solution if only two directions are\nallowed. This result sheds new light on the structure of optimal solutions,\nwhich may help guide algorithm design for the general problem.\n", "date": "2023-05-25", "categories": []} {"id": "2305.14632", "title": "Supermodular Rank: Set Function Decomposition and Optimization", "abstract": " We define the supermodular rank of a function on a lattice. This is the\nsmallest number of terms needed to decompose it into a sum of supermodular\nfunctions. The supermodular summands are defined with respect to different\npartial orders. We characterize the maximum possible value of the supermodular\nrank and describe the functions with fixed supermodular rank. We analogously\ndefine the submodular rank. We use submodular decompositions to optimize set\nfunctions. Given a bound on the submodular rank of a set function, we formulate\nan algorithm that splits an optimization problem into submodular subproblems.\nWe show that this method improves the approximation ratio guarantees of several\nalgorithms for monotone set function maximization and ratio of set functions\nminimization, at a computation overhead that depends on the submodular rank.\n", "date": "2023-05-24", "categories": []} {"id": "2208.03860", "title": "Rankability and Linear Ordering Problem: New Probabilistic Insight and\n Algorithms", "abstract": " The linear ordering problem (LOP), which consists in ordering M objects from\ntheir pairwise comparisons, is commonly applied in many areas of research.\nWhile efforts have been made to devise efficient LOP algorithms, verification\nof whether the data are rankable, that is, if the linear ordering problem (LOP)\nsolutions have a meaningful interpretation, received much less attention. To\naddress this problem, we adopt a probabilistic perspective where the results of\npairwise comparisons are modeled as Bernoulli variables with a common parameter\nand we estimate the latter from the observed data. The brute-force approach to\nthe required enumeration has a prohibitive complexity of O(M !), so we\nreformulate the problem and introduce a concept of the Slater spectrum that\ngeneralizes the Slater index, and then devise an algorithm to find the spectrum\nwith complexity O(M^3 2^M) that is manageable for moderate values of M.\nFurthermore, with a minor modification of the algorithm, we are able to find\nall solutions of the LOP with the complexity O(M 2^M). Numerical examples are\nshown on synthetic and real-world data, and the algorithms are publicly\navailable.\n", "date": "2022-08-08", "categories": []} {"id": "2206.13181", "title": "Learning to Control Local Search for Combinatorial Optimization", "abstract": " Combinatorial optimization problems are encountered in many practical\ncontexts such as logistics and production, but exact solutions are particularly\ndifficult to find and usually NP-hard for considerable problem sizes. To\ncompute approximate solutions, a zoo of generic as well as problem-specific\nvariants of local search is commonly used. However, which variant to apply to\nwhich particular problem is difficult to decide even for experts.\n In this paper we identify three independent algorithmic aspects of such local\nsearch algorithms and formalize their sequential selection over an optimization\nprocess as Markov Decision Process (MDP). We design a deep graph neural network\nas policy model for this MDP, yielding a learned controller for local search\ncalled NeuroLS. Ample experimental evidence shows that NeuroLS is able to\noutperform both, well-known general purpose local search controllers from\nOperations Research as well as latest machine learning-based approaches.\n", "date": "2022-06-27", "categories": []} {"id": "2302.05134", "title": "Neural Capacitated Clustering", "abstract": " Recent work on deep clustering has found new promising methods also for\nconstrained clustering problems. Their typically pairwise constraints often can\nbe used to guide the partitioning of the data. Many problems however, feature\ncluster-level constraints, e.g. the Capacitated Clustering Problem (CCP), where\neach point has a weight and the total weight sum of all points in each cluster\nis bounded by a prescribed capacity. In this paper we propose a new method for\nthe CCP, Neural Capacited Clustering, that learns a neural network to predict\nthe assignment probabilities of points to cluster centers from a data set of\noptimal or near optimal past solutions of other problem instances. During\ninference, the resulting scores are then used in an iterative k-means like\nprocedure to refine the assignment under capacity constraints. In our\nexperiments on artificial data and two real world datasets our approach\noutperforms several state-of-the-art mathematical and heuristic solvers from\nthe literature. Moreover, we apply our method in the context of a\ncluster-first-route-second approach to the Capacitated Vehicle Routing Problem\n(CVRP) and show competitive results on the well-known Uchoa benchmark.\n", "date": "2023-02-10", "categories": []} {"id": "2305.11648", "title": "Applying Ising Machines to Multi-objective QUBOs", "abstract": " Multi-objective optimisation problems involve finding solutions with varying\ntrade-offs between multiple and often conflicting objectives. Ising machines\nare physical devices that aim to find the absolute or approximate ground states\nof an Ising model. To apply Ising machines to multi-objective problems, a\nweighted sum objective function is used to convert multi-objective into\nsingle-objective problems. However, deriving scalarisation weights that\narchives evenly distributed solutions across the Pareto front is not trivial.\nPrevious work has shown that adaptive weights based on dichotomic search, and\none based on averages of previously explored weights can explore the Pareto\nfront quicker than uniformly generated weights. However, these adaptive methods\nhave only been applied to bi-objective problems in the past. In this work, we\nextend the adaptive method based on averages in two ways: (i)~we extend the\nadaptive method of deriving scalarisation weights for problems with two or more\nobjectives, and (ii)~we use an alternative measure of distance to improve\nperformance.\n We compare the proposed method with existing ones and show that it leads to\nthe best performance on multi-objective Unconstrained Binary Quadratic\nProgramming (mUBQP) instances with 3 and 4 objectives and that it is\ncompetitive with the best one for instances with 2 objectives.\n", "date": "2023-05-19", "categories": []} {"id": "2305.10444", "title": "Optimizing Forest Fire Prevention: Intelligent Scheduling Algorithms for\n Drone-Based Surveillance System", "abstract": " Given the importance of forests and their role in maintaining the ecological\nbalance, which directly affects the planet, the climate, and the life on this\nplanet, this research presents the problem of forest fire monitoring using\ndrones. The forest monitoring process is performed continuously to track any\nchanges in the monitored region within the forest. During fires, drones'\ncapture data is used to increase the follow-up speed and enhance the control\nprocess of these fires to prevent their spread. The time factor in such\nproblems determines the success rate of the fire extinguishing process, as\nappropriate data at the right time may be the decisive factor in controlling\nfires, preventing their spread, extinguishing them, and limiting their losses.\nTherefore, this research presented the problem of monitoring task scheduling\nfor drones in the forest monitoring system. This problem is solved by\ndeveloping several algorithms with the aim of minimizing the total completion\ntime required to carry out all the drones' assigned tasks. System performance\nis measured by using 990 instances of three different classes. The performed\nexperimental results indicated the effectiveness of the proposed algorithms and\ntheir ability to act efficiently to achieve the desired goal. The algorithm\n$RID$ achieved the best performance with a percentage rate of up to 90.3% with\na time of 0.088 seconds.\n", "date": "2023-05-14", "categories": []} {"id": "2305.10192", "title": "Curriculum Learning in Job Shop Scheduling using Reinforcement Learning", "abstract": " Solving job shop scheduling problems (JSSPs) with a fixed strategy, such as a\npriority dispatching rule, may yield satisfactory results for several problem\ninstances but, nevertheless, insufficient results for others. From this\nsingle-strategy perspective finding a near optimal solution to a specific JSSP\nvaries in difficulty even if the machine setup remains the same. A recent\nintensively researched and promising method to deal with difficulty variability\nis Deep Reinforcement Learning (DRL), which dynamically adjusts an agent's\nplanning strategy in response to difficult instances not only during training,\nbut also when applied to new situations. In this paper, we further improve DLR\nas an underlying method by actively incorporating the variability of difficulty\nwithin the same problem size into the design of the learning process. We base\nour approach on a state-of-the-art methodology that solves JSSP by means of DRL\nand graph neural network embeddings. Our work supplements the training routine\nof the agent by a curriculum learning strategy that ranks the problem instances\nshown during training by a new metric of problem instance difficulty. Our\nresults show that certain curricula lead to significantly better performances\nof the DRL solutions. Agents trained on these curricula beat the top\nperformance of those trained on randomly distributed training data, reaching\n3.2% shorter average makespans.\n", "date": "2023-05-17", "categories": []} {"id": "2305.09257", "title": "A new node-shift encoding representation for the travelling salesman\n problem", "abstract": " This paper presents a new genetic algorithm encoding representation to solve\nthe travelling salesman problem. To assess the performance of the proposed\nchromosome structure, we compare it with state-of-the-art encoding\nrepresentations. For that purpose, we use 14 benchmarks of different sizes\ntaken from TSPLIB. Finally, after conducting the experimental study, we report\nthe obtained results and draw our conclusion.\n", "date": "2023-05-16", "categories": []} {"id": "2305.07178", "title": "Fast Pareto Optimization Using Sliding Window Selection", "abstract": " Pareto optimization using evolutionary multi-objective algorithms has been\nwidely applied to solve constrained submodular optimization problems. A crucial\nfactor determining the runtime of the used evolutionary algorithms to obtain\ngood approximations is the population size of the algorithms which grows with\nthe number of trade-offs that the algorithms encounter. In this paper, we\nintroduce a sliding window speed up technique for recently introduced\nalgorithms. We prove that our technique eliminates the population size as a\ncrucial factor negatively impacting the runtime and achieves the same\ntheoretical performance guarantees as previous approaches within less\ncomputation time. Our experimental investigations for the classical maximum\ncoverage problem confirms that our sliding window technique clearly leads to\nbetter results for a wide range of instances and constraint settings.\n", "date": "2023-05-11", "categories": []} {"id": "2305.03574", "title": "Scope Restriction for Scalable Real-Time Railway Rescheduling: An\n Exploratory Study", "abstract": " With the aim to stimulate future research, we describe an exploratory study\nof a railway rescheduling problem. A widely used approach in practice and state\nof the art is to decompose these complex problems by geographical scope.\nInstead, we propose defining a core problem that restricts a rescheduling\nproblem in response to a disturbance to only trains that need to be\nrescheduled, hence restricting the scope in both time and space. In this\ncontext, the difficulty resides in defining a scoper that can predict a subset\nof train services that will be affected by a given disturbance. We report\npreliminary results using the Flatland simulation environment that highlights\nthe potential and challenges of this idea. We provide an extensible playground\nopen-source implementation based on the Flatland railway environment and\nAnswer-Set Programming.\n", "date": "2023-05-05", "categories": []} {"id": "2305.03632", "title": "Improving LaCAM for Scalable Eventually Optimal Multi-Agent Pathfinding", "abstract": " This study extends the recently-developed LaCAM algorithm for multi-agent\npathfinding (MAPF). LaCAM is a sub-optimal search-based algorithm that uses\nlazy successor generation to dramatically reduce the planning effort. We\npresent two enhancements. First, we propose its anytime version, called LaCAM*,\nwhich eventually converges to optima, provided that solution costs are\naccumulated transition costs. Second, we improve the successor generation to\nquickly obtain initial solutions. Exhaustive experiments demonstrate their\nutility. For instance, LaCAM* sub-optimally solved 99% of the instances\nretrieved from the MAPF benchmark, where the number of agents varied up to a\nthousand, within ten seconds on a standard desktop PC, while ensuring eventual\nconvergence to optima; developing a new horizon of MAPF algorithms.\n", "date": "2023-05-05", "categories": []} {"id": "2305.02805", "title": "Local Optima Correlation Assisted Adaptive Operator Selection", "abstract": " For solving combinatorial optimisation problems with metaheuristics,\ndifferent search operators are applied for sampling new solutions in the\nneighbourhood of a given solution. It is important to understand the\nrelationship between operators for various purposes, e.g., adaptively deciding\nwhen to use which operator to find optimal solutions efficiently. However, it\nis difficult to theoretically analyse this relationship, especially in the\ncomplex solution space of combinatorial optimisation problems. In this paper,\nwe propose to empirically analyse the relationship between operators in terms\nof the correlation between their local optima and develop a measure for\nquantifying their relationship. The comprehensive analyses on a wide range of\ncapacitated vehicle routing problem benchmark instances show that there is a\nconsistent pattern in the correlation between commonly used operators. Based on\nthis newly proposed local optima correlation metric, we propose a novel\napproach for adaptively selecting among the operators during the search\nprocess. The core intention is to improve search efficiency by preventing\nwasting computational resources on exploring neighbourhoods where the local\noptima have already been reached. Experiments on randomly generated instances\nand commonly used benchmark datasets are conducted. Results show that the\nproposed approach outperforms commonly used adaptive operator selection\nmethods.\n", "date": "2023-05-03", "categories": []} {"id": "2305.01942", "title": "Experimental Design for Any $p$-Norm", "abstract": " We consider a general $p$-norm objective for experimental design problems\nthat captures some well-studied objectives (D/A/E-design) as special cases. We\nprove that a randomized local search approach provides a unified algorithm to\nsolve this problem for all $p$. This provides the first approximation algorithm\nfor the general $p$-norm objective, and a nice interpolation of the best known\nbounds of the special cases.\n", "date": "2023-05-03", "categories": []} {"id": "2207.02057", "title": "Online 2-stage Stable Matching", "abstract": " We focus on an online 2-stage problem, motivated by the following situation:\nconsider a system where students shall be assigned to universities. There is a\nfirst round where some students apply, and a first (stable) matching $M_1$ has\nto be computed. However, some students may decide to leave the system (change\ntheir plan, go to a foreign university, or to some institution not in the\nsystem). Then, in a second round (after these deletions), we shall compute a\nsecond (final) stable matching $M_2$. As it is undesirable to change\nassignments, the goal is to minimize the number of divorces/modifications\nbetween the two stable matchings $M_1$ and $M_2$. Then, how should we choose\n$M_1$ and $M_2$? We show that there is an {\\it optimal online} algorithm to\nsolve this problem. In particular, thanks to a dominance property, we show that\nwe can optimally compute $M_1$ without knowing the students that will leave the\nsystem. We generalize the result to some other possible modifications in the\ninput (students, open positions).\n We also tackle the case of more stages, showing that no competitive (online)\nalgorithm can be achieved for the considered problem as soon as there are 3\nstages.\n", "date": "2022-07-05", "categories": []} {"id": "2305.01623", "title": "Augmented Electronic Ising Machine as an Effective SAT Solver", "abstract": " With the slowdown of improvement in conventional von Neumann systems,\nincreasing attention is paid to novel paradigms such as Ising machines. They\nhave very different approach to NP-complete optimization problems. Ising\nmachines have shown great potential in solving binary optimization problems\nlike MaxCut. In this paper, we present an analysis of these systems in\nsatisfiability (SAT) problems. We demonstrate that, in the case of 3-SAT, a\nbasic architecture fails to produce meaningful acceleration, thanks in no small\npart to the relentless progress made in conventional SAT solvers. Nevertheless,\ncareful analysis attributes part of the failure to the lack of two important\ncomponents: cubic interactions and efficient randomization heuristics. To\novercome these limitations, we add proper architectural support for cubic\ninteraction on a state-of-the-art Ising machine. More importantly, we propose a\nnovel semantic-aware annealing schedule that makes the search-space navigation\nmuch more efficient than existing annealing heuristics. With experimental\nanalyses, we show that such an Augmented Ising Machine for SAT (AIMS),\noutperforms state-of-the-art software-based, GPU-based and conventional\nhardware SAT solvers by orders of magnitude. We also demonstrate AIMS to be\nrelatively robust against device variation and noise.\n", "date": "2023-05-01", "categories": []} {"id": "2305.00535", "title": "Nearly Optimal Steiner Trees using Graph Neural Network Assisted Monte\n Carlo Tree Search", "abstract": " Graph neural networks are useful for learning problems, as well as for\ncombinatorial and graph problems such as the Subgraph Isomorphism Problem and\nthe Traveling Salesman Problem. We describe an approach for computing Steiner\nTrees by combining a graph neural network and Monte Carlo Tree Search. We first\ntrain a graph neural network that takes as input a partial solution and\nproposes a new node to be added as output. This neural network is then used in\na Monte Carlo search to compute a Steiner tree. The proposed method\nconsistently outperforms the standard 2-approximation algorithm on many\ndifferent types of graphs and often finds the optimal solution.\n", "date": "2023-04-30", "categories": []} {"id": "2301.02053", "title": "Max-Min Diversification with Fairness Constraints: Exact and\n Approximation Algorithms", "abstract": " Diversity maximization aims to select a diverse and representative subset of\nitems from a large dataset. It is a fundamental optimization task that finds\napplications in data summarization, feature selection, web search, recommender\nsystems, and elsewhere. However, in a setting where data items are associated\nwith different groups according to sensitive attributes like sex or race, it is\npossible that algorithmic solutions for this task, if left unchecked, will\nunder- or over-represent some of the groups. Therefore, we are motivated to\naddress the problem of \\emph{max-min diversification with fairness\nconstraints}, aiming to select $k$ items to maximize the minimum distance\nbetween any pair of selected items while ensuring that the number of items\nselected from each group falls within predefined lower and upper bounds. In\nthis work, we propose an exact algorithm based on integer linear programming\nthat is suitable for small datasets as well as a\n$\\frac{1-\\varepsilon}{5}$-approximation algorithm for any $\\varepsilon \\in (0,\n1)$ that scales to large datasets. Extensive experiments on real-world datasets\ndemonstrate the superior performance of our proposed algorithms over existing\nones.\n", "date": "2023-01-05", "categories": []} {"id": "2304.09975", "title": "Solving the Kidney-Exchange Problem via Graph Neural Networks with No\n Supervision", "abstract": " This paper introduces a new learning-based approach for approximately solving\nthe Kidney-Exchange Problem (KEP), an NP-hard problem on graphs. The problem\nconsists of, given a pool of kidney donors and patients waiting for kidney\ndonations, optimally selecting a set of donations to optimize the quantity and\nquality of transplants performed while respecting a set of constraints about\nthe arrangement of these donations. The proposed technique consists of two main\nsteps: the first is a Graph Neural Network (GNN) trained without supervision;\nthe second is a deterministic non-learned search heuristic that uses the output\nof the GNN to find paths and cycles. To allow for comparisons, we also\nimplemented and tested an exact solution method using integer programming, two\ngreedy search heuristics without the machine learning module, and the GNN alone\nwithout a heuristic. We analyze and compare the methods and conclude that the\nlearning-based two-stage approach is the best solution quality, outputting\napproximate solutions on average 1.1 times more valuable than the ones from the\ndeterministic heuristic alone.\n", "date": "2023-04-19", "categories": []} {"id": "2304.09395", "title": "H-TSP: Hierarchically Solving the Large-Scale Travelling Salesman\n Problem", "abstract": " We propose an end-to-end learning framework based on hierarchical\nreinforcement learning, called H-TSP, for addressing the large-scale Travelling\nSalesman Problem (TSP). The proposed H-TSP constructs a solution of a TSP\ninstance starting from the scratch relying on two components: the upper-level\npolicy chooses a small subset of nodes (up to 200 in our experiment) from all\nnodes that are to be traversed, while the lower-level policy takes the chosen\nnodes as input and outputs a tour connecting them to the existing partial route\n(initially only containing the depot). After jointly training the upper-level\nand lower-level policies, our approach can directly generate solutions for the\ngiven TSP instances without relying on any time-consuming search procedures. To\ndemonstrate effectiveness of the proposed approach, we have conducted extensive\nexperiments on randomly generated TSP instances with different numbers of\nnodes. We show that H-TSP can achieve comparable results (gap 3.42% vs. 7.32%)\nas SOTA search-based approaches, and more importantly, we reduce the time\nconsumption up to two orders of magnitude (3.32s vs. 395.85s). To the best of\nour knowledge, H-TSP is the first end-to-end deep reinforcement learning\napproach that can scale to TSP instances of up to 10000 nodes. Although there\nare still gaps to SOTA results with respect to solution quality, we believe\nthat H-TSP will be useful for practical applications, particularly those that\nare time-sensitive e.g., on-call routing and ride hailing service.\n", "date": "2023-04-19", "categories": []} {"id": "2304.09407", "title": "Pointerformer: Deep Reinforced Multi-Pointer Transformer for the\n Traveling Salesman Problem", "abstract": " Traveling Salesman Problem (TSP), as a classic routing optimization problem\noriginally arising in the domain of transportation and logistics, has become a\ncritical task in broader domains, such as manufacturing and biology. Recently,\nDeep Reinforcement Learning (DRL) has been increasingly employed to solve TSP\ndue to its high inference efficiency. Nevertheless, most of existing end-to-end\nDRL algorithms only perform well on small TSP instances and can hardly\ngeneralize to large scale because of the drastically soaring memory consumption\nand computation time along with the enlarging problem scale. In this paper, we\npropose a novel end-to-end DRL approach, referred to as Pointerformer, based on\nmulti-pointer Transformer. Particularly, Pointerformer adopts both reversible\nresidual network in the encoder and multi-pointer network in the decoder to\neffectively contain memory consumption of the encoder-decoder architecture. To\nfurther improve the performance of TSP solutions, Pointerformer employs both a\nfeature augmentation method to explore the symmetries of TSP at both training\nand inference stages as well as an enhanced context embedding approach to\ninclude more comprehensive context information in the query. Extensive\nexperiments on a randomly generated benchmark and a public benchmark have shown\nthat, while achieving comparative results on most small-scale TSP instances as\nSOTA DRL approaches do, Pointerformer can also well generalize to large-scale\nTSPs.\n", "date": "2023-04-19", "categories": []} {"id": "2304.06048", "title": "RELS-DQN: A Robust and Efficient Local Search Framework for\n Combinatorial Optimization", "abstract": " Combinatorial optimization (CO) aims to efficiently find the best solution to\nNP-hard problems ranging from statistical physics to social media marketing. A\nwide range of CO applications can benefit from local search methods because\nthey allow reversible action over greedy policies. Deep Q-learning (DQN) using\nmessage-passing neural networks (MPNN) has shown promise in replicating the\nlocal search behavior and obtaining comparable results to the local search\nalgorithms. However, the over-smoothing and the information loss during the\niterations of message passing limit its robustness across applications, and the\nlarge message vectors result in memory inefficiency. Our paper introduces\nRELS-DQN, a lightweight DQN framework that exhibits the local search behavior\nwhile providing practical scalability. Using the RELS-DQN model trained on one\napplication, it can generalize to various applications by providing solution\nvalues higher than or equal to both the local search algorithms and the\nexisting DQN models while remaining efficient in runtime and memory.\n", "date": "2023-04-11", "categories": []} {"id": "2206.05576", "title": "Optimal Solutions for Joint Beamforming and Antenna Selection: From\n Branch and Bound to Graph Neural Imitation Learning", "abstract": " This work revisits the joint beamforming (BF) and antenna selection (AS)\nproblem, as well as its robust beamforming (RBF) version under imperfect\nchannel state information (CSI). Such problems arise due to various reasons,\ne.g., the costly nature of the radio frequency (RF) chains and\nenergy/resource-saving considerations. The joint (R)BF\\&AS problem is a mixed\ninteger and nonlinear program, and thus finding {\\it optimal solutions} is\noften costly, if not outright impossible. The vast majority of the prior works\ntackled these problems using techniques such as continuous approximations,\ngreedy methods, and supervised machine learning -- yet these approaches do not\nensure optimality or even feasibility of the solutions. The main contribution\nof this work is threefold. First, an effective {\\it branch and bound} (B\\&B)\nframework for solving the problems of interest is proposed. Leveraging existing\nBF and RBF solvers, it is shown that the B\\&B framework guarantees global\noptimality of the considered problems. Second, to expedite the potentially\ncostly B\\&B algorithm, a machine learning (ML)-based scheme is proposed to help\nskip intermediate states of the B\\&B search tree. The learning model features a\n{\\it graph neural network} (GNN)-based design that is resilient to a commonly\nencountered challenge in wireless communications, namely, the change of problem\nsize (e.g., the number of users) across the training and test stages. Third,\ncomprehensive performance characterizations are presented, showing that the\nGNN-based method retains the global optimality of B\\&B with provably reduced\ncomplexity, under reasonable conditions. Numerical simulations also show that\nthe ML-based acceleration can often achieve an order-of-magnitude speedup\nrelative to B\\&B.\n", "date": "2022-06-11", "categories": []} {"id": "2304.04022", "title": "A Reinforcement Learning-assisted Genetic Programming Algorithm for Team\n Formation Problem Considering Person-Job Matching", "abstract": " An efficient team is essential for the company to successfully complete new\nprojects. To solve the team formation problem considering person-job matching\n(TFP-PJM), a 0-1 integer programming model is constructed, which considers both\nperson-job matching and team members' willingness to communicate on team\nefficiency, with the person-job matching score calculated using intuitionistic\nfuzzy numbers. Then, a reinforcement learning-assisted genetic programming\nalgorithm (RL-GP) is proposed to enhance the quality of solutions. The RL-GP\nadopts the ensemble population strategies. Before the population evolution at\neach generation, the agent selects one from four population search modes\naccording to the information obtained, thus realizing a sound balance of\nexploration and exploitation. In addition, surrogate models are used in the\nalgorithm to evaluate the formation plans generated by individuals, which\nspeeds up the algorithm learning process. Afterward, a series of comparison\nexperiments are conducted to verify the overall performance of RL-GP and the\neffectiveness of the improved strategies within the algorithm. The\nhyper-heuristic rules obtained through efficient learning can be utilized as\ndecision-making aids when forming project teams. This study reveals the\nadvantages of reinforcement learning methods, ensemble strategies, and the\nsurrogate model applied to the GP framework. The diversity and intelligent\nselection of search patterns along with fast adaptation evaluation, are\ndistinct features that enable RL-GP to be deployed in real-world enterprise\nenvironments.\n", "date": "2023-04-08", "categories": []} {"id": "2304.04051", "title": "Generating a Graph Colouring Heuristic with Deep Q-Learning and Graph\n Neural Networks", "abstract": " The graph colouring problem consists of assigning labels, or colours, to the\nvertices of a graph such that no two adjacent vertices share the same colour.\nIn this work we investigate whether deep reinforcement learning can be used to\ndiscover a competitive construction heuristic for graph colouring. Our proposed\napproach, ReLCol, uses deep Q-learning together with a graph neural network for\nfeature extraction, and employs a novel way of parameterising the graph that\nresults in improved performance. Using standard benchmark graphs with varied\ntopologies, we empirically evaluate the benefits and limitations of the\nheuristic learned by ReLCol relative to existing construction algorithms, and\ndemonstrate that reinforcement learning is a promising direction for further\nresearch on the graph colouring problem.\n", "date": "2023-04-08", "categories": []} {"id": "2304.04217", "title": "The Study of Highway for Lifelong Multi-Agent Path Finding", "abstract": " In modern fulfillment warehouses, agents traverse the map to complete endless\ntasks that arrive on the fly, which is formulated as a lifelong Multi-Agent\nPath Finding (lifelong MAPF) problem. The goal of tackling this challenging\nproblem is to find the path for each agent in a finite runtime while maximizing\nthe throughput. However, existing methods encounter exponential growth of\nruntime and undesirable phenomena of deadlocks and rerouting as the map size or\nagent density grows. To address these challenges in lifelong MAPF, we explore\nthe idea of highways mainly studied for one-shot MAPF (i.e., finding paths at\nonce beforehand), which reduces the complexity of the problem by encouraging\nagents to move in the same direction. We utilize two methods to incorporate the\nhighway idea into the lifelong MAPF framework and discuss the properties that\nminimize the existing problems of deadlocks and rerouting. The experimental\nresults demonstrate that the runtime is considerably reduced and the decay of\nthroughput is gradually insignificant as the map size enlarges under the\nsettings of the highway. Furthermore, when the density of agents increases, the\nphenomena of deadlocks and rerouting are significantly reduced by leveraging\nthe highway.\n", "date": "2023-04-09", "categories": []} {"id": "2304.04700", "title": "Achieving Long-term Fairness in Submodular Maximization through\n Randomization", "abstract": " Submodular function optimization has numerous applications in machine\nlearning and data analysis, including data summarization which aims to identify\na concise and diverse set of data points from a large dataset. It is important\nto implement fairness-aware algorithms when dealing with data items that may\ncontain sensitive attributes like race or gender, to prevent biases that could\nlead to unequal representation of different groups. With this in mind, we\ninvestigate the problem of maximizing a monotone submodular function while\nmeeting group fairness constraints. Unlike previous studies in this area, we\nallow for randomized solutions, with the objective being to calculate a\ndistribution over feasible sets such that the expected number of items selected\nfrom each group is subject to constraints in the form of upper and lower\nthresholds, ensuring that the representation of each group remains balanced in\nthe long term. Here a set is considered feasible if its size does not exceed a\nconstant value of $b$. Our research includes the development of a series of\napproximation algorithms for this problem.\n", "date": "2023-04-10", "categories": []} {"id": "2304.03755", "title": "Online Learning for Scheduling MIP Heuristics", "abstract": " Mixed Integer Programming (MIP) is NP-hard, and yet modern solvers often\nsolve large real-world problems within minutes. This success can partially be\nattributed to heuristics. Since their behavior is highly instance-dependent,\nrelying on hard-coded rules derived from empirical testing on a large\nheterogeneous corpora of benchmark instances might lead to sub-optimal\nperformance. In this work, we propose an online learning approach that adapts\nthe application of heuristics towards the single instance at hand. We replace\nthe commonly used static heuristic handling with an adaptive framework\nexploiting past observations about the heuristic's behavior to make future\ndecisions. In particular, we model the problem of controlling Large\nNeighborhood Search and Diving - two broad and complex classes of heuristics -\nas a multi-armed bandit problem. Going beyond existing work in the literature,\nwe control two different classes of heuristics simultaneously by a single\nlearning agent. We verify our approach numerically and show consistent node\nreductions over the MIPLIB 2017 Benchmark set. For harder instances that take\nat least 1000 seconds to solve, we observe a speedup of 4%.\n", "date": "2023-04-04", "categories": []} {"id": "2303.06531", "title": "Towards Practical Multi-Robot Hybrid Tasks Allocation for Autonomous\n Cleaning", "abstract": " Task allocation plays a vital role in multi-robot autonomous cleaning\nsystems, where multiple robots work together to clean a large area. However,\nmost current studies mainly focus on deterministic, single-task allocation for\ncleaning robots, without considering hybrid tasks in uncertain working\nenvironments. Moreover, there is a lack of datasets and benchmarks for relevant\nresearch. In this paper, to address these problems, we formulate multi-robot\nhybrid-task allocation under the uncertain cleaning environment as a robust\noptimization problem. Firstly, we propose a novel robust mixed-integer linear\nprogramming model with practical constraints including the task order\nconstraint for different tasks and the ability constraints of hybrid robots.\nSecondly, we establish a dataset of \\emph{100} instances made from floor plans,\neach of which has 2D manually-labeled images and a 3D model. Thirdly, we\nprovide comprehensive results on the collected dataset using three traditional\noptimization approaches and a deep reinforcement learning-based solver. The\nevaluation results show that our solution meets the needs of multi-robot\ncleaning task allocation and the robust solver can protect the system from\nworst-case scenarios with little additional cost. The benchmark will be\navailable at\n{https://github.com/iamwangyabin/Multi-robot-Cleaning-Task-Allocation}.\n", "date": "2023-03-12", "categories": []} {"id": "2302.04694", "title": "Partial Optimality in Cubic Correlation Clustering", "abstract": " The higher-order correlation clustering problem is an expressive model, and\nrecently, local search heuristics have been proposed for several applications.\nCertifying optimality, however, is NP-hard and practically hampered already by\nthe complexity of the problem statement. Here, we focus on establishing partial\noptimality conditions for the special case of complete graphs and cubic\nobjective functions. In addition, we define and implement algorithms for\ntesting these conditions and examine their effect numerically, on two datasets.\n", "date": "2023-02-09", "categories": []} {"id": "2303.14793", "title": "Multi-Phase Relaxation Labeling for Square Jigsaw Puzzle Solving", "abstract": " We present a novel method for solving square jigsaw puzzles based on global\noptimization. The method is fully automatic, assumes no prior information, and\ncan handle puzzles with known or unknown piece orientation. At the core of the\noptimization process is nonlinear relaxation labeling, a well-founded approach\nfor deducing global solutions from local constraints, but unlike the classical\nscheme here we propose a multi-phase approach that guarantees convergence to\nfeasible puzzle solutions. Next to the algorithmic novelty, we also present a\nnew compatibility function for the quantification of the affinity between\nadjacent puzzle pieces. Competitive results and the advantage of the\nmulti-phase approach are demonstrated on standard datasets.\n", "date": "2023-03-26", "categories": []} {"id": "2303.13532", "title": "Enhanced Iterated local search for the technician routing and scheduling\n problem", "abstract": " Most public facilities in the European countries, including France, Germany,\nand the UK, were built during the reconstruction projects between 1950 and\n1980. Owing to the deteriorating state of such vital infrastructure has become\nrelatively expensive in the recent decades. A significant part of the\nmaintenance operation costs is spent on the technical staff. Therefore, the\noptimal use of the available workforce is essential to optimize the operation\ncosts. This includes planning technical interventions, workload balancing,\nproductivity improvement, etc. In this paper, we focus on the routing of\ntechnicians and scheduling of their tasks. We address for this purpose a\nvariant of the workforce scheduling problem called the technician routing and\nscheduling problem (TRSP). This problem has applications in different fields,\nsuch as transportation infrastructure (rail and road networks),\ntelecommunications, and sewage facilities. To solve the TRSP, we propose an\nenhanced iterated local search (eILS) approach. The enhancement of the ILS\nfirstly includes an intensification procedure that incorporates a set of local\nsearch operators and removal-repair heuristics crafted for the TRSP. Next, four\ndifferent mechanisms are used in the perturbation phase. Finally, an elite set\nof solutions is used to extensively explore the neighborhood of local optima as\nwell as to enhance diversification during search space exploration. To measure\nthe performance of the proposed method, experiments were conducted based on\nbenchmark instances from the literature, and the results obtained were compared\nwith those of an existing method. Our method achieved very good results, since\nit reached the best overall gap, which is three times lower than that of the\nliterature. Furthermore, eILS improved the best-known solution for $34$\ninstances among a total of $56$ while maintaining reasonable computational\ntimes.\n", "date": "2023-03-12", "categories": []} {"id": "2303.13117", "title": "RLOR: A Flexible Framework of Deep Reinforcement Learning for Operation\n Research", "abstract": " Reinforcement learning has been applied in operation research and has shown\npromise in solving large combinatorial optimization problems. However, existing\nworks focus on developing neural network architectures for certain problems.\nThese works lack the flexibility to incorporate recent advances in\nreinforcement learning, as well as the flexibility of customizing model\narchitectures for operation research problems. In this work, we analyze the\nend-to-end autoregressive models for vehicle routing problems and show that\nthese models can benefit from the recent advances in reinforcement learning\nwith a careful re-implementation of the model architecture. In particular, we\nre-implemented the Attention Model and trained it with Proximal Policy\nOptimization (PPO) in CleanRL, showing at least 8 times speed up in training\ntime. We hereby introduce RLOR, a flexible framework for Deep Reinforcement\nLearning for Operation Research. We believe that a flexible framework is key to\ndeveloping deep reinforcement learning models for operation research problems.\nThe code of our work is publicly available at https://github.com/cpwan/RLOR.\n", "date": "2023-03-23", "categories": []} {"id": "2303.09960", "title": "Stochastic Submodular Maximization via Polynomial Estimators", "abstract": " In this paper, we study stochastic submodular maximization problems with\ngeneral matroid constraints, that naturally arise in online learning, team\nformation, facility location, influence maximization, active learning and\nsensing objective functions. In other words, we focus on maximizing submodular\nfunctions that are defined as expectations over a class of submodular functions\nwith an unknown distribution. We show that for monotone functions of this form,\nthe stochastic continuous greedy algorithm attains an approximation ratio (in\nexpectation) arbitrarily close to $(1-1/e) \\approx 63\\%$ using a polynomial\nestimation of the gradient. We argue that using this polynomial estimator\ninstead of the prior art that uses sampling eliminates a source of randomness\nand experimentally reduces execution time.\n", "date": "2023-03-17", "categories": []} {"id": "2303.08219", "title": "A 2-opt Algorithm for Locally Optimal Set Partition Optimization", "abstract": " Our research deals with the optimization version of the set partition\nproblem, where the objective is to minimize the absolute difference between the\nsums of the two disjoint partitions. Although this problem is known to be\nNP-hard and requires exponential time to solve, we propose a less demanding\nversion of this problem where the goal is to find a locally optimal solution.\nIn our approach, we consider the local optimality in respect to any movement of\nat most two elements. To accomplish this, we developed an algorithm that can\ngenerate a locally optimal solution in at most $O(N^2)$ time and $O(N)$ space.\nOur algorithm can handle arbitrary input precisions and does not require\npositive or integer inputs. Hence, it can be applied in various problem\nscenarios with ease.\n", "date": "2023-03-14", "categories": []} {"id": "2303.07603", "title": "Redrawing attendance boundaries to promote racial and ethnic diversity\n in elementary schools", "abstract": " Most US school districts draw \"attendance boundaries\" to define catchment\nareas that assign students to schools near their homes, often recapitulating\nneighborhood demographic segregation in schools. Focusing on elementary\nschools, we ask: how much might we reduce school segregation by redrawing\nattendance boundaries? Combining parent preference data with methods from\ncombinatorial optimization, we simulate alternative boundaries for 98 US school\ndistricts serving over 3 million elementary-aged students, minimizing\nWhite/non-White segregation while mitigating changes to travel times and school\nsizes. Across districts, we observe a median 14% relative decrease in\nsegregation, which we estimate would require approximately 20\\% of students to\nswitch schools and, surprisingly, a slight reduction in travel times. We\nrelease a public dashboard depicting these alternative boundaries\n(https://www.schooldiversity.org/) and invite both school boards and their\nconstituents to evaluate their viability. Our results show the possibility of\ngreater integration without significant disruptions for families.\n", "date": "2023-03-14", "categories": []} {"id": "2208.14698", "title": "Bayesian Optimization-based Combinatorial Assignment", "abstract": " We study the combinatorial assignment domain, which includes combinatorial\nauctions and course allocation. The main challenge in this domain is that the\nbundle space grows exponentially in the number of items. To address this,\nseveral papers have recently proposed machine learning-based preference\nelicitation algorithms that aim to elicit only the most important information\nfrom agents. However, the main shortcoming of this prior work is that it does\nnot model a mechanism's uncertainty over values for not yet elicited bundles.\nIn this paper, we address this shortcoming by presenting a Bayesian\noptimization-based combinatorial assignment (BOCA) mechanism. Our key technical\ncontribution is to integrate a method for capturing model uncertainty into an\niterative combinatorial auction mechanism. Concretely, we design a new method\nfor estimating an upper uncertainty bound that can be used to define an\nacquisition function to determine the next query to the agents. This enables\nthe mechanism to properly explore (and not just exploit) the bundle space\nduring its preference elicitation phase. We run computational experiments in\nseveral spectrum auction domains to evaluate BOCA's performance. Our results\nshow that BOCA achieves higher allocative efficiency than state-of-the-art\napproaches.\n", "date": "2022-08-31", "categories": []} {"id": "2302.06926", "title": "Lightsolver challenges a leading deep learning solver for Max-2-SAT\n problems", "abstract": " Maximum 2-satisfiability (MAX-2-SAT) is a type of combinatorial decision\nproblem that is known to be NP-hard. In this paper, we compare LightSolver's\nquantum-inspired algorithm to a leading deep-learning solver for the MAX-2-SAT\nproblem. Experiments on benchmark data sets show that LightSolver achieves\nsignificantly smaller time-to-optimal-solution compared to a state-of-the-art\ndeep-learning algorithm, where the gain in performance tends to increase with\nthe problem size.\n", "date": "2023-02-14", "categories": []} {"id": "2303.07202", "title": "Optimization of the location and design of urban green spaces", "abstract": " The recent promotion of sustainable urban planning combined with a growing\nneed for public interventions to improve well-being and health have led to an\nincreased collective interest for green spaces in and around cities. In\nparticular, parks have proven a wide range of benefits in urban areas. This\nalso means inequities in park accessibility may contribute to health\ninequities. In this work, we showcase the application of classic tools from\nOperations Research to assist decision-makers to improve parks' accessibility,\ndistribution and design. Given the context of public decision-making, we are\nparticularly concerned with equity and environmental justice, and are focused\non an advanced assessment of users' behavior through a spatial interaction\nmodel. We present a two-stage fair facility location and design model, which\nserves as a template model to assist public decision-makers at the city-level\nfor the planning of urban green spaces. The first-stage of the optimization\nmodel is about the optimal city-budget allocation to neighborhoods based on a\ndata exposing inequality attributes. The second-stage seeks the optimal\nlocation and design of parks for each neighborhood, and the objective consists\nof maximizing the total expected probability of individuals visiting parks. We\nshow how to reformulate the latter as a mixed-integer linear program. We\nfurther introduce a clustering method to reduce the size of the problem and\ndetermine a close to optimal solution within reasonable time. The model is\ntested using the case study of the city of Montreal and comparative results are\ndiscussed in detail to justify the performance of the model.\n", "date": "2023-03-13", "categories": []} {"id": "2302.01546", "title": "Group Fairness in Non-monotone Submodular Maximization", "abstract": " Maximizing a submodular function has a wide range of applications in machine\nlearning and data mining. One such application is data summarization whose goal\nis to select a small set of representative and diverse data items from a large\ndataset. However, data items might have sensitive attributes such as race or\ngender, in this setting, it is important to design \\emph{fairness-aware}\nalgorithms to mitigate potential algorithmic bias that may cause over- or\nunder- representation of particular groups. Motivated by that, we propose and\nstudy the classic non-monotone submodular maximization problem subject to novel\ngroup fairness constraints. Our goal is to select a set of items that maximizes\na non-monotone submodular function, while ensuring that the number of selected\nitems from each group is proportionate to its size, to the extent specified by\nthe decision maker. We develop the first constant-factor approximation\nalgorithms for this problem. We also extend the basic model to incorporate an\nadditional global size constraint on the total number of selected items.\n", "date": "2023-02-03", "categories": []} {"id": "2303.04147", "title": "Heuristics for Vehicle Routing Problem: A Survey and Recent Advances", "abstract": " Vehicle routing is a well-known optimization research topic with significant\npractical importance. Among different approaches to solving vehicle routing,\nheuristics can produce a satisfactory solution at a reasonable computational\ncost. Consequently, much effort has been made in the past decades to develop\nvehicle routing heuristics. In this article, we systematically survey the\nexisting vehicle routing heuristics, particularly on works carried out in\nrecent years. A classification of vehicle routing heuristics is presented,\nfollowed by a review of their methodologies, recent developments, and\napplications. Moreover, we present a general framework of state-of-the-art\nmethods and provide insights into their success. Finally, three emerging\nresearch topics with notable works and future directions are discussed.\n", "date": "2023-03-01", "categories": []} {"id": "2303.04333", "title": "Preference-Aware Delivery Planning for Last-Mile Logistics", "abstract": " Optimizing delivery routes for last-mile logistics service is challenging and\nhas attracted the attention of many researchers. These problems are usually\nmodeled and solved as variants of vehicle routing problems (VRPs) with\nchallenging real-world constraints (e.g., time windows, precedence). However,\ndespite many decades of solid research on solving these VRP instances, we still\nsee significant gaps between optimized routes and the routes that are actually\npreferred by the practitioners. Most of these gaps are due to the difference\nbetween what's being optimized, and what the practitioners actually care about,\nwhich is hard to be defined exactly in many instances. In this paper, we\npropose a novel hierarchical route optimizer with learnable parameters that\ncombines the strength of both the optimization and machine learning approaches.\nOur hierarchical router first solves a zone-level Traveling Salesman Problem\nwith learnable weights on various zone-level features; with the zone visit\nsequence fixed, we then solve the stop-level vehicle routing problem as a\nShortest Hamiltonian Path problem. The Bayesian optimization approach is then\nintroduced to allow us to adjust the weights to be assigned to different zone\nfeatures used in solving the zone-level Traveling Salesman Problem. By using a\nreal-world delivery dataset provided by the Amazon Last Mile Routing Research\nChallenge, we demonstrate the importance of having both the optimization and\nthe machine learning components. We also demonstrate how we can use\nroute-related features to identify instances that we might have difficulty\nwith. This paves ways to further research on how we can tackle these difficult\ninstances.\n", "date": "2023-03-08", "categories": []} {"id": "2303.03475", "title": "Rolling Horizon based Temporal Decomposition for the Offline Pickup and\n Delivery Problem with Time Windows", "abstract": " The offline pickup and delivery problem with time windows (PDPTW) is a\nclassical combinatorial optimization problem in the transportation community,\nwhich has proven to be very challenging computationally. Due to the complexity\nof the problem, practical problem instances can be solved only via heuristics,\nwhich trade-off solution quality for computational tractability. Among the\nvarious heuristics, a common strategy is problem decomposition, that is, the\nreduction of a large-scale problem into a collection of smaller sub-problems,\nwith spatial and temporal decompositions being two natural approaches. While\nspatial decomposition has been successful in certain settings, effective\ntemporal decomposition has been challenging due to the difficulty of stitching\ntogether the sub-problem solutions across the decomposition boundaries. In this\nwork, we introduce a novel temporal decomposition scheme for solving a class of\nPDPTWs that have narrow time windows, for which it is able to provide both fast\nand high-quality solutions. We utilize techniques that have been popularized\nrecently in the context of online dial-a-ride problems along with the general\nidea of rolling horizon optimization. To the best of our knowledge, this is the\nfirst attempt to solve offline PDPTWs using such an approach. To show the\nperformance and scalability of our framework, we use the optimization of\nparatransit services as a motivating example. We compare our results with an\noffline heuristic algorithm using Google OR-Tools. In smaller problem\ninstances, the baseline approach is as competitive as our framework. However,\nin larger problem instances, our framework is more scalable and can provide\ngood solutions to problem instances of varying degrees of difficulty, while the\nbaseline algorithm often fails to find a feasible solution within comparable\ncompute times.\n", "date": "2023-03-06", "categories": []} {"id": "2206.09895", "title": "Multiple Fairness and Cardinality constraints for Students-Topics\n Grouping Problem", "abstract": " Group work is a prevalent activity in educational settings, where students\nare often divided into topic-specific groups based on their preferences. The\ngrouping should reflect the students' aspirations as much as possible. Usually,\nthe resulting groups should also be balanced in terms of protected attributes\nlike gender or race since studies indicate that students might learn better in\na diverse group. Moreover, balancing the group cardinalities is also an\nessential requirement for fair workload distribution across the groups. In this\npaper, we introduce the multi-fair capacitated (MFC) grouping problem that\nfairly partitions students into non-overlapping groups while ensuring balanced\ngroup cardinalities (with a lower bound and an upper bound), and maximizing the\ndiversity of members in terms of protected attributes. We propose two\napproaches: a heuristic method and a knapsack-based method to obtain the MFC\ngrouping. The experiments on a real dataset and a semi-synthetic dataset show\nthat our proposed methods can satisfy students' preferences well and deliver\nbalanced and diverse groups regarding cardinality and the protected attribute,\nrespectively.\n", "date": "2022-06-20", "categories": []} {"id": "2301.10703", "title": "A Sequential Deep Learning Algorithm for Sampled Mixed-integer\n Optimisation Problems", "abstract": " Mixed-integer optimisation problems can be computationally challenging. Here,\nwe introduce and analyse two efficient algorithms with a specific sequential\ndesign that are aimed at dealing with sampled problems within this class. At\neach iteration step of both algorithms, we first test the feasibility of a\ngiven test solution for each and every constraint associated with the sampled\noptimisation at hand, while also identifying those constraints that are\nviolated. Subsequently, an optimisation problem is constructed with a\nconstraint set consisting of the current basis -- namely, the smallest set of\nconstraints that fully specifies the current test solution -- as well as\nconstraints related to a limited number of the identified violating samples. We\nshow that both algorithms exhibit finite-time convergence towards the optimal\nsolution. Algorithm 2 features a neural network classifier that notably\nimproves the computational performance compared to Algorithm 1. We\nquantitatively establish these algorithms' efficacy through three numerical\ntests: robust optimal power flow, robust unit commitment, and robust random\nmixed-integer linear program.\n", "date": "2023-01-25", "categories": []} {"id": "2302.13941", "title": "A Reinforcement Learning Approach for Scheduling Problems With Improved\n Generalization Through Order Swapping", "abstract": " The scheduling of production resources (such as associating jobs to machines)\nplays a vital role for the manufacturing industry not only for saving energy\nbut also for increasing the overall efficiency. Among the different job\nscheduling problems, the JSSP is addressed in this work. JSSP falls into the\ncategory of NP-hard COP, in which solving the problem through exhaustive search\nbecomes unfeasible. Simple heuristics such as FIFO, LPT and metaheuristics such\nas Taboo search are often adopted to solve the problem by truncating the search\nspace. The viability of the methods becomes inefficient for large problem sizes\nas it is either far from the optimum or time consuming. In recent years, the\nresearch towards using DRL to solve COP has gained interest and has shown\npromising results in terms of solution quality and computational efficiency. In\nthis work, we provide an novel approach to solve the JSSP examining the\nobjectives generalization and solution effectiveness using DRL. In particular,\nwe employ the PPO algorithm that adopts the policy-gradient paradigm that is\nfound to perform well in the constrained dispatching of jobs. We incorporated\nan OSM in the environment to achieve better generalized learning of the\nproblem. The performance of the presented approach is analyzed in depth by\nusing a set of available benchmark instances and comparing our results with the\nwork of other groups.\n", "date": "2023-02-27", "categories": []} {"id": "2303.02442", "title": "Neural Airport Ground Handling", "abstract": " Airport ground handling (AGH) offers necessary operations to flights during\ntheir turnarounds and is of great importance to the efficiency of airport\nmanagement and the economics of aviation. Such a problem involves the interplay\namong the operations that leads to NP-hard problems with complex constraints.\nHence, existing methods for AGH are usually designed with massive domain\nknowledge but still fail to yield high-quality solutions efficiently. In this\npaper, we aim to enhance the solution quality and computation efficiency for\nsolving AGH. Particularly, we first model AGH as a multiple-fleet vehicle\nrouting problem (VRP) with miscellaneous constraints including precedence, time\nwindows, and capacity. Then we propose a construction framework that decomposes\nAGH into sub-problems (i.e., VRPs) in fleets and present a neural method to\nconstruct the routing solutions to these sub-problems. In specific, we resort\nto deep learning and parameterize the construction heuristic policy with an\nattention-based neural network trained with reinforcement learning, which is\nshared across all sub-problems. Extensive experiments demonstrate that our\nmethod significantly outperforms classic meta-heuristics, construction\nheuristics and the specialized methods for AGH. Besides, we empirically verify\nthat our neural method generalizes well to instances with large numbers of\nflights or varying parameters, and can be readily adapted to solve real-time\nAGH with stochastic flight arrivals. Our code is publicly available at:\nhttps://github.com/RoyalSkye/AGH.\n", "date": "2023-03-04", "categories": []} {"id": "2303.01774", "title": "Bayesian Optimization over High-Dimensional Combinatorial Spaces via\n Dictionary-based Embeddings", "abstract": " We consider the problem of optimizing expensive black-box functions over\nhigh-dimensional combinatorial spaces which arises in many science,\nengineering, and ML applications. We use Bayesian Optimization (BO) and propose\na novel surrogate modeling approach for efficiently handling a large number of\nbinary and categorical parameters. The key idea is to select a number of\ndiscrete structures from the input space (the dictionary) and use them to\ndefine an ordinal embedding for high-dimensional combinatorial structures. This\nallows us to use existing Gaussian process models for continuous spaces. We\ndevelop a principled approach based on binary wavelets to construct\ndictionaries for binary spaces, and propose a randomized construction method\nthat generalizes to categorical spaces. We provide theoretical justification to\nsupport the effectiveness of the dictionary-based embeddings. Our experiments\non diverse real-world benchmarks demonstrate the effectiveness of our proposed\nsurrogate modeling approach over state-of-the-art BO methods.\n", "date": "2023-03-03", "categories": []} {"id": "2303.01963", "title": "Multi-Start Team Orienteering Problem for UAS Mission Re-Planning with\n Data-Efficient Deep Reinforcement Learning", "abstract": " In this paper, we study the Multi-Start Team Orienteering Problem (MSTOP), a\nmission re-planning problem where vehicles are initially located away from the\ndepot and have different amounts of fuel. We consider/assume the goal of\nmultiple vehicles is to travel to maximize the sum of collected profits under\nresource (e.g., time, fuel) consumption constraints. Such re-planning problems\noccur in a wide range of intelligent UAS applications where changes in the\nmission environment force the operation of multiple vehicles to change from the\noriginal plan. To solve this problem with deep reinforcement learning (RL), we\ndevelop a policy network with self-attention on each partial tour and\nencoder-decoder attention between the partial tour and the remaining nodes. We\npropose a modified REINFORCE algorithm where the greedy rollout baseline is\nreplaced by a local mini-batch baseline based on multiple, possibly\nnon-duplicate sample rollouts. By drawing multiple samples per training\ninstance, we can learn faster and obtain a stable policy gradient estimator\nwith significantly fewer instances. The proposed training algorithm outperforms\nthe conventional greedy rollout baseline, even when combined with the maximum\nentropy objective.\n", "date": "2023-03-02", "categories": []} {"id": "2303.00466", "title": "ASP: Learn a Universal Neural Solver!", "abstract": " Applying machine learning to combinatorial optimization problems has the\npotential to improve both efficiency and accuracy. However, existing\nlearning-based solvers often struggle with generalization when faced with\nchanges in problem distributions and scales. In this paper, we propose a new\napproach called ASP: Adaptive Staircase Policy Space Response Oracle to address\nthese generalization issues and learn a universal neural solver. ASP consists\nof two components: Distributional Exploration, which enhances the solver's\nability to handle unknown distributions using Policy Space Response Oracles,\nand Persistent Scale Adaption, which improves scalability through curriculum\nlearning. We have tested ASP on several challenging COPs, including the\ntraveling salesman problem, the vehicle routing problem, and the prize\ncollecting TSP, as well as the real-world instances from TSPLib and CVRPLib.\nOur results show that even with the same model size and weak training signal,\nASP can help neural solvers explore and adapt to unseen distributions and\nvarying scales, achieving superior performance. In particular, compared with\nthe same neural solvers under a standard training pipeline, ASP produces a\nremarkable decrease in terms of the optimality gap with 90.9% and 47.43% on\ngenerated instances and real-world instances for TSP, and a decrease of 19% and\n45.57% for CVRP.\n", "date": "2023-03-01", "categories": []} {"id": "2302.13797", "title": "Learning Large Neighborhood Search for Vehicle Routing in Airport Ground\n Handling", "abstract": " Dispatching vehicle fleets to serve flights is a key task in airport ground\nhandling (AGH). Due to the notable growth of flights, it is challenging to\nsimultaneously schedule multiple types of operations (services) for a large\nnumber of flights, where each type of operation is performed by one specific\nvehicle fleet. To tackle this issue, we first represent the operation\nscheduling as a complex vehicle routing problem and formulate it as a mixed\ninteger linear programming (MILP) model. Then given the graph representation of\nthe MILP model, we propose a learning assisted large neighborhood search (LNS)\nmethod using data generated based on real scenarios, where we integrate\nimitation learning and graph convolutional network (GCN) to learn a destroy\noperator to automatically select variables, and employ an off-the-shelf solver\nas the repair operator to reoptimize the selected variables. Experimental\nresults based on a real airport show that the proposed method allows for\nhandling up to 200 flights with 10 types of operations simultaneously, and\noutperforms state-of-the-art methods. Moreover, the learned method performs\nconsistently accompanying different solvers, and generalizes well on larger\ninstances, verifying the versatility and scalability of our method.\n", "date": "2023-02-27", "categories": []} {"id": "2302.14442", "title": "City-scale Pollution Aware Traffic Routing by Sampling Max Flows using\n MCMC", "abstract": " A significant cause of air pollution in urban areas worldwide is the high\nvolume of road traffic. Long-term exposure to severe pollution can cause\nserious health issues. One approach towards tackling this problem is to design\na pollution-aware traffic routing policy that balances multiple objectives of\ni) avoiding extreme pollution in any area ii) enabling short transit times, and\niii) making effective use of the road capacities. We propose a novel\nsampling-based approach for this problem. We provide the first construction of\na Markov Chain that can sample integer max flow solutions of a planar graph,\nwith theoretical guarantees that the probabilities depend on the aggregate\ntransit length. We designed a traffic policy using diverse samples and\nsimulated traffic on real-world road maps using the SUMO traffic simulator. We\nobserve a considerable decrease in areas with severe pollution when\nexperimented with maps of large cities across the world compared to other\napproaches.\n", "date": "2023-02-28", "categories": []} {"id": "2302.14678", "title": "Graph Reinforcement Learning for Operator Selection in the ALNS\n Metaheuristic", "abstract": " ALNS is a popular metaheuristic with renowned efficiency in solving\ncombinatorial optimisation problems. However, despite 16 years of intensive\nresearch into ALNS, whether the embedded adaptive layer can efficiently select\noperators to improve the incumbent remains an open question. In this work, we\nformulate the choice of operators as a Markov Decision Process, and propose a\npractical approach based on Deep Reinforcement Learning and Graph Neural\nNetworks. The results show that our proposed method achieves better performance\nthan the classic ALNS adaptive layer due to the choice of operator being\nconditioned on the current solution. We also discuss important considerations\nsuch as the size of the operator portfolio and the impact of the choice of\noperator scales. Notably, our approach can also save significant time and\nlabour costs for handcrafting problem-specific operator portfolios.\n", "date": "2023-02-28", "categories": []} {"id": "2302.13225", "title": "Towards Tackling MaxSAT by Combining Nested Monte Carlo with Local\n Search", "abstract": " Recent work proposed the UCTMAXSAT algorithm to address Maximum\nSatisfiability Problems (MaxSAT) and shown improved performance over pure\nStochastic Local Search algorithms (SLS). UCTMAXSAT is based on Monte Carlo\nTree Search but it uses SLS instead of purely random playouts. In this work, we\nintroduce two algorithmic variations over UCTMAXSAT. We carry an empirical\nanalysis on MaxSAT benchmarks from recent competitions and establish that both\nideas lead to performance improvements. First, a nesting of the tree search\ninspired by the Nested Monte Carlo Search algorithm is effective on most\ninstance types in the benchmark. Second, we observe that using a static flip\nlimit in SLS, the ideal budget depends heavily on the instance size and we\npropose to set it dynamically. We show that it is a robust way to achieve\ncomparable performance on a variety of instances without requiring additional\ntuning.\n", "date": "2023-02-26", "categories": []} {"id": "2302.11992", "title": "Solving Recurrent MIPs with Semi-supervised Graph Neural Networks", "abstract": " We propose an ML-based model that automates and expedites the solution of\nMIPs by predicting the values of variables. Our approach is motivated by the\nobservation that many problem instances share salient features and solution\nstructures since they differ only in few (time-varying) parameters. Examples\ninclude transportation and routing problems where decisions need to be\nre-optimized whenever commodity volumes or link costs change. Our method is the\nfirst to exploit the sequential nature of the instances being solved\nperiodically, and can be trained with ``unlabeled'' instances, when exact\nsolutions are unavailable, in a semi-supervised setting. Also, we provide a\nprincipled way of transforming the probabilistic predictions into integral\nsolutions. Using a battery of experiments with representative binary MIPs, we\nshow the gains of our model over other ML-based optimization approaches.\n", "date": "2023-02-20", "categories": []} {"id": "2302.05446", "title": "Robust Scheduling with GFlowNets", "abstract": " Finding the best way to schedule operations in a computation graph is a\nclassical NP-hard problem which is central to compiler optimization. However,\nevaluating the goodness of a schedule on the target hardware can be very\ntime-consuming. Traditional approaches as well as previous machine learning\nones typically optimize proxy metrics, which are fast to evaluate but can lead\nto bad schedules when tested on the target hardware. In this work, we propose a\nnew approach to scheduling by sampling proportionally to the proxy metric using\na novel GFlowNet method. We introduce a technique to control the trade-off\nbetween diversity and goodness of the proposed schedules at inference time and\ndemonstrate empirically that the pure optimization baselines can lead to subpar\nperformance with respect to our approach when tested on a target model.\nFurthermore, we show that conditioning the GFlowNet on the computation graph\nenables generalization to unseen scheduling problems for both synthetic and\nreal-world compiler datasets.\n", "date": "2023-01-17", "categories": []} {"id": "2302.07162", "title": "Semiconductor Fab Scheduling with Self-Supervised and Reinforcement\n Learning", "abstract": " Semiconductor manufacturing is a notoriously complex and costly multi-step\nprocess involving a long sequence of operations on expensive and\nquantity-limited equipment. Recent chip shortages and their impacts have\nhighlighted the importance of semiconductors in the global supply chains and\nhow reliant on those our daily lives are. Due to the investment cost,\nenvironmental impact, and time scale needed to build new factories, it is\ndifficult to ramp up production when demand spikes.\n This work introduces a method to successfully learn to schedule a\nsemiconductor manufacturing facility more efficiently using deep reinforcement\nand self-supervised learning. We propose the first adaptive scheduling approach\nto handle complex, continuous, stochastic, dynamic, modern semiconductor\nmanufacturing models. Our method outperforms the traditional hierarchical\ndispatching strategies typically used in semiconductor manufacturing plants,\nsubstantially reducing each order's tardiness and time until completion. As a\nresult, our method yields a better allocation of resources in the semiconductor\nmanufacturing process.\n", "date": "2023-02-14", "categories": []} {"id": "2210.13694", "title": "Worst-Case Adaptive Submodular Cover", "abstract": " In this paper, we study the adaptive submodular cover problem under the\nworst-case setting. This problem generalizes many previously studied problems,\nnamely, the pool-based active learning and the stochastic submodular set cover.\nThe input of our problem is a set of items (e.g., medical tests) and each item\nhas a random state (e.g., the outcome of a medical test), whose realization is\ninitially unknown. One must select an item at a fixed cost in order to observe\nits realization. There is an utility function which maps a subset of items and\ntheir states to a non-negative real number. We aim to sequentially select a\ngroup of items to achieve a ``target value'' while minimizing the maximum cost\nacross realizations (a.k.a. worst-case cost). To facilitate our study, we\nassume that the utility function is \\emph{worst-case submodular}, a property\nthat is commonly found in many machine learning applications. With this\nassumption, we develop a tight $(\\log (Q/\\eta)+1)$-approximation policy, where\n$Q$ is the ``target value'' and $\\eta$ is the smallest difference between $Q$\nand any achievable utility value $\\hat{Q}\nis proposed. The RL-GA also uses an elite individual retention strategy to\nimprove search performance. After that, a task time window selection algorithm\n(TTWSA) is proposed to evaluate the performance of population evolution.\nSeveral experiments are used to examine the scheduling effect of the proposed\nalgorithm. Through the experimental verification of multiple instances, it can\nbe seen that the RL-GA can solve the EDSSP problem effectively. Compared with\nthe state-of-the-art algorithms, the RL-GA performs better in several aspects.\n", "date": "2022-06-12", "categories": []} {"id": "2301.02142", "title": "Playing hide and seek: tackling in-store picking operations while\n improving customer experience", "abstract": " The evolution of the retail business presents new challenges and raises\npivotal questions on how to reinvent stores and supply chains to meet the\ngrowing demand of the online channel. One of the recent measures adopted by\nomnichannel retailers is to address the growth of online sales using in-store\npicking, which allows serving online orders using existing assets. However, it\ncomes with the downside of harming the offline customer experience. To achieve\npicking policies adapted to the dynamic customer flows of a retail store, we\nformalize a new problem called Dynamic In-store Picker Routing Problem (diPRP).\nIn this relevant problem - diPRP - a picker tries to pick online orders while\nminimizing customer encounters. We model the problem as a Markov Decision\nProcess (MDP) and solve it using a hybrid solution approach comprising\nmathematical programming and reinforcement learning components. Computational\nexperiments on synthetic instances suggest that the algorithm converges to\nefficient policies. Furthermore, we apply our approach in the context of a\nlarge European retailer to assess the results of the proposed policies\nregarding the number of orders picked and customers encountered. Our work\nsuggests that retailers should be able to scale the in-store picking of online\norders without jeopardizing the experience of offline customers. The policies\nlearned using the proposed solution approach reduced the number of customer\nencounters by more than 50% when compared to policies solely focused on picking\norders. Thus, to pursue omnichannel strategies that adequately trade-off\noperational efficiency and customer experience, retailers cannot rely on actual\nsimplistic picking strategies, such as choosing the shortest possible route.\n", "date": "2023-01-05", "categories": []} {"id": "2206.13211", "title": "Modern graph neural networks do worse than classical greedy algorithms\n in solving combinatorial optimization problems like maximum independent set", "abstract": " The recent work ``Combinatorial Optimization with Physics-Inspired Graph\nNeural Networks'' [Nat Mach Intell 4 (2022) 367] introduces a physics-inspired\nunsupervised Graph Neural Network (GNN) to solve combinatorial optimization\nproblems on sparse graphs. To test the performances of these GNNs, the authors\nof the work show numerical results for two fundamental problems: maximum cut\nand maximum independent set (MIS). They conclude that \"the graph neural network\noptimizer performs on par or outperforms existing solvers, with the ability to\nscale beyond the state of the art to problems with millions of variables.\"\n In this comment, we show that a simple greedy algorithm, running in almost\nlinear time, can find solutions for the MIS problem of much better quality than\nthe GNN. The greedy algorithm is faster by a factor of $10^4$ with respect to\nthe GNN for problems with a million variables. We do not see any good reason\nfor solving the MIS with these GNN, as well as for using a sledgehammer to\ncrack nuts.\n In general, many claims of superiority of neural networks in solving\ncombinatorial problems are at risk of being not solid enough, since we lack\nstandard benchmarks based on really hard problems. We propose one of such hard\nbenchmarks, and we hope to see future neural network optimizers tested on these\nproblems before any claim of superiority is made.\n", "date": "2022-06-27", "categories": []} {"id": "2210.00623", "title": "Inability of a graph neural network heuristic to outperform greedy\n algorithms in solving combinatorial optimization problems like Max-Cut", "abstract": " In Nature Machine Intelligence 4, 367 (2022), Schuetz et al provide a scheme\nto employ graph neural networks (GNN) as a heuristic to solve a variety of\nclassical, NP-hard combinatorial optimization problems. It describes how the\nnetwork is trained on sample instances and the resulting GNN heuristic is\nevaluated applying widely used techniques to determine its ability to succeed.\nClearly, the idea of harnessing the powerful abilities of such networks to\n``learn'' the intricacies of complex, multimodal energy landscapes in such a\nhands-off approach seems enticing. And based on the observed performance, the\nheuristic promises to be highly scalable, with a computational cost linear in\nthe input size $n$, although there is likely a significant overhead in the\npre-factor due to the GNN itself. However, closer inspection shows that the\nreported results for this GNN are only minutely better than those for gradient\ndescent and get outperformed by a greedy algorithm, for example, for Max-Cut.\nThe discussion also highlights what I believe are some common misconceptions in\nthe evaluations of heuristics.\n", "date": "2022-10-02", "categories": []} {"id": "2206.14987", "title": "Lookback for Learning to Branch", "abstract": " The expressive and computationally inexpensive bipartite Graph Neural\nNetworks (GNN) have been shown to be an important component of deep learning\nbased Mixed-Integer Linear Program (MILP) solvers. Recent works have\ndemonstrated the effectiveness of such GNNs in replacing the branching\n(variable selection) heuristic in branch-and-bound (B&B) solvers. These GNNs\nare trained, offline and on a collection of MILPs, to imitate a very good but\ncomputationally expensive branching heuristic, strong branching. Given that B&B\nresults in a tree of sub-MILPs, we ask (a) whether there are strong\ndependencies exhibited by the target heuristic among the neighboring nodes of\nthe B&B tree, and (b) if so, whether we can incorporate them in our training\nprocedure. Specifically, we find that with the strong branching heuristic, a\nchild node's best choice was often the parent's second-best choice. We call\nthis the \"lookback\" phenomenon. Surprisingly, the typical branching GNN of\nGasse et al. (2019) often misses this simple \"answer\". To imitate the target\nbehavior more closely by incorporating the lookback phenomenon in GNNs, we\npropose two methods: (a) target smoothing for the standard cross-entropy loss\nfunction, and (b) adding a Parent-as-Target (PAT) Lookback regularizer term.\nFinally, we propose a model selection framework to incorporate\nharder-to-formulate objectives such as solving time in the final models.\nThrough extensive experimentation on standard benchmark instances, we show that\nour proposal results in up to 22% decrease in the size of the B&B tree and up\nto 15% improvement in the solving times.\n", "date": "2022-06-30", "categories": []} {"id": "2212.11451", "title": "A machine learning framework for neighbor generation in metaheuristic\n search", "abstract": " This paper presents a methodology for integrating machine learning techniques\ninto metaheuristics for solving combinatorial optimization problems. Namely, we\npropose a general machine learning framework for neighbor generation in\nmetaheuristic search. We first define an efficient neighborhood structure\nconstructed by applying a transformation to a selected subset of variables from\nthe current solution. Then, the key of the proposed methodology is to generate\npromising neighbors by selecting a proper subset of variables that contains a\ndescent of the objective in the solution space. To learn a good variable\nselection strategy, we formulate the problem as a classification task that\nexploits structural information from the characteristics of the problem and\nfrom high-quality solutions. We validate our methodology on two metaheuristic\napplications: a Tabu Search scheme for solving a Wireless Network Optimization\nproblem and a Large Neighborhood Search heuristic for solving Mixed-Integer\nPrograms. The experimental results show that our approach is able to achieve a\nsatisfactory trade-off between the exploration of a larger solution space and\nthe exploitation of high-quality solution regions on both applications.\n", "date": "2022-12-22", "categories": []} {"id": "2212.11730", "title": "TransPath: Learning Heuristics For Grid-Based Pathfinding via\n Transformers", "abstract": " Heuristic search algorithms, e.g. A*, are the commonly used tools for\npathfinding on grids, i.e. graphs of regular structure that are widely employed\nto represent environments in robotics, video games etc. Instance-independent\nheuristics for grid graphs, e.g. Manhattan distance, do not take the obstacles\ninto account and, thus, the search led by such heuristics performs poorly in\nthe obstacle-rich environments. To this end, we suggest learning the\ninstance-dependent heuristic proxies that are supposed to notably increase the\nefficiency of the search. The first heuristic proxy we suggest to learn is the\ncorrection factor, i.e. the ratio between the instance independent cost-to-go\nestimate and the perfect one (computed offline at the training phase). Unlike\nlearning the absolute values of the cost-to-go heuristic function, which was\nknown before, when learning the correction factor the knowledge of the\ninstance-independent heuristic is utilized. The second heuristic proxy is the\npath probability, which indicates how likely the grid cell is lying on the\nshortest path. This heuristic can be utilized in the Focal Search framework as\nthe secondary heuristic, allowing us to preserve the guarantees on the bounded\nsub-optimality of the solution. We learn both suggested heuristics in a\nsupervised fashion with the state-of-the-art neural networks containing\nattention blocks (transformers). We conduct a thorough empirical evaluation on\na comprehensive dataset of planning tasks, showing that the suggested\ntechniques i) reduce the computational effort of the A* up to a factor of $4$x\nwhile producing the solutions, which costs exceed the costs of the optimal\nsolutions by less than $0.3$% on average; ii) outperform the competitors, which\ninclude the conventional techniques from the heuristic search, i.e. weighted\nA*, as well as the state-of-the-art learnable planners.\n", "date": "2022-12-22", "categories": []} {"id": "2212.11011", "title": "Hidden-Variables Genetic Algorithm for Variable-Size Design Space\n Optimal Layout Problems with Application to Aerospace Vehicles", "abstract": " The optimal layout of a complex system such as aerospace vehicles consists in\nplacing a given number of components in a container in order to minimize one or\nseveral objectives under some geometrical or functional constraints. This paper\npresents an extended formulation of this problem as a variable-size design\nspace (VSDS) problem to take into account a large number of architectural\nchoices and components allocation during the design process. As a\nrepresentative example of such systems, considering the layout of a satellite\nmodule, the VSDS aspect translates the fact that the optimizer has to choose\nbetween several subdivisions of the components. For instance, one large tank of\nfuel might be placed as well as two smaller tanks or three even smaller tanks\nfor the same amount of fuel. In order to tackle this NP-hard problem, a genetic\nalgorithm enhanced by an adapted hidden-variables mechanism is proposed. This\nlatter is illustrated on a toy case and an aerospace application case\nrepresentative to real world complexity to illustrate the performance of the\nproposed algorithms. The results obtained using the proposed mechanism are\nreported and analyzed.\n", "date": "2022-12-21", "categories": []} {"id": "2212.08101", "title": "Learning to repeatedly solve routing problems", "abstract": " In the last years, there has been a great interest in machine-learning-based\nheuristics for solving NP-hard combinatorial optimization problems. The\ndeveloped methods have shown potential on many optimization problems. In this\npaper, we present a learned heuristic for the reoptimization of a problem after\na minor change in its data. We focus on the case of the capacited vehicle\nrouting problem with static clients (i.e., same client locations) and changed\ndemands. Given the edges of an original solution, the goal is to predict and\nfix the ones that have a high chance of remaining in an optimal solution after\na change of client demands. This partial prediction of the solution reduces the\ncomplexity of the problem and speeds up its resolution, while yielding a good\nquality solution. The proposed approach resulted in solutions with an\noptimality gap ranging from 0\\% to 1.7\\% on different benchmark instances\nwithin a reasonable computing time.\n", "date": "2022-12-15", "categories": []} {"id": "2212.08299", "title": "Metaheuristic for Hub-Spoke Facility Location Problem: Application to\n Indian E-commerce Industry", "abstract": " Indian e-commerce industry has evolved over the last decade and is expected\nto grow over the next few years. The focus has now shifted to turnaround time\n(TAT) due to the emergence of many third-party logistics providers and higher\ncustomer expectations. The key consideration for delivery providers is to\nbalance their overall operating costs while meeting the promised TAT to their\ncustomers. E-commerce delivery partners operate through a network of facilities\nwhose strategic locations help to run the operations efficiently. In this work,\nwe identify the locations of hubs throughout the country and their\ncorresponding mapping with the distribution centers. The objective is to\nminimize the total network costs with TAT adherence. We use Genetic Algorithm\nand leverage business constraints to reduce the solution search space and hence\nthe solution time. The results indicate an improvement of 9.73% in TAT\ncompliance compared with the current scenario.\n", "date": "2022-12-16", "categories": []} {"id": "2212.07543", "title": "Monte-Carlo Tree-Search for Leveraging Performance of Blackbox Job-Shop\n Scheduling Heuristics", "abstract": " In manufacturing, the production is often done on out-of-the-shelf\nmanufacturing lines, whose underlying scheduling heuristics are not known due\nto the intellectual property. We consider such a setting with a black-box\njob-shop system and an unknown scheduling heuristic that, for a given\npermutation of jobs, schedules the jobs for the black-box job-shop with the\ngoal of minimizing the makespan. Here, the jobs need to enter the job-shop in\nthe given order of the permutation, but may take different paths within the job\nshop, which depends on the black-box heuristic. The performance of the\nblack-box heuristic depends on the order of the jobs, and the natural problem\nfor the manufacturer is to find an optimum ordering of the jobs.\n Facing a real-world scenario as described above, we engineer the Monte-Carlo\ntree-search for finding a close-to-optimum ordering of jobs. To cope with a\nlarge solutions-space in planning scenarios, a hierarchical Monte-Carlo tree\nsearch (H-MCTS) is proposed based on abstraction of jobs. On synthetic and\nreal-life problems, H-MCTS with integrated abstraction significantly\noutperforms pure heuristic-based techniques as well as other Monte-Carlo search\nvariants. We furthermore show that, by modifying the evaluation metric in\nH-MCTS, it is possible to achieve other optimization objectives than what the\nscheduling heuristics are designed for -- e.g., minimizing the total completion\ntime instead of the makespan. Our experimental observations have been also\nvalidated in real-life cases, and our H-MCTS approach has been implemented in a\nproduction plant's controller.\n", "date": "2022-12-14", "categories": []} {"id": "2212.06940", "title": "Heuristically Guided Compilation for Multi-Agent Path Finding", "abstract": " Multi-agent path finding (MAPF) is a task of finding non-conflicting paths\nconnecting agents' specified initial and goal positions in a shared\nenvironment. We focus on compilation-based solvers in which the MAPF problem is\nexpressed in a different well established formalism such as mixed-integer\nlinear programming (MILP), Boolean satisfiability (SAT), or constraint\nprogramming (CP). As the target solvers for these formalisms act as black-boxes\nit is challenging to integrate MAPF specific heuristics in the MAPF\ncompilation-based solvers. We show in this work how the build a MAPF encoding\nfor the target SAT solver in which domain specific heuristic knowledge is\nreflected. The heuristic knowledge is transferred to the SAT solver by\nselecting candidate paths for each agent and by constructing the encoding only\nfor these candidate paths instead of constructing the encoding for all possible\npaths for an agent. The conducted experiments show that heuristically guided\ncompilation outperforms the vanilla variants of the SAT-based MAPF solver.\n", "date": "2022-12-13", "categories": []} {"id": "2210.12933", "title": "Multi-Agent Path Finding via Tree LSTM", "abstract": " In recent years, Multi-Agent Path Finding (MAPF) has attracted attention from\nthe fields of both Operations Research (OR) and Reinforcement Learning (RL).\nHowever, in the 2021 Flatland3 Challenge, a competition on MAPF, the best RL\nmethod scored only 27.9, far less than the best OR method. This paper proposes\na new RL solution to Flatland3 Challenge, which scores 125.3, several times\nhigher than the best RL solution before. We creatively apply a novel network\narchitecture, TreeLSTM, to MAPF in our solution. Together with several other RL\ntechniques, including reward shaping, multiple-phase training, and centralized\ncontrol, our solution is comparable to the top 2-3 OR methods.\n", "date": "2022-10-24", "categories": []} {"id": "2207.12840", "title": "Partial-Monotone Adaptive Submodular Maximization", "abstract": " Many sequential decision making problems, including pool-based active\nlearning and adaptive viral marketing, can be formulated as an adaptive\nsubmodular maximization problem. Most of existing studies on adaptive\nsubmodular optimization focus on either monotone case or non-monotone case.\nSpecifically, if the utility function is monotone and adaptive submodular,\n\\cite{golovin2011adaptive} developed a greedy policy that achieves a $(1-1/e)$\napproximation ratio subject to a cardinality constraint. If the utility\nfunction is non-monotone and adaptive submodular, \\cite{tang2021beyond} showed\nthat a random greedy policy achieves a $1/e$ approximation ratio subject to a\ncardinality constraint. In this work, we aim to generalize the above mentioned\nresults by studying the partial-monotone adaptive submodular maximization\nproblem. To this end, we introduce the notation of adaptive monotonicity ratio\n$m\\in[0,1]$ to measure the degree of monotonicity of a function. Our main\nresult is to show that a random greedy policy achieves an approximation ratio\nof $m(1-1/e)+(1-m)(1/e)$ if the utility function is $m$-adaptive monotone and\nadaptive submodular. Notably this result recovers the aforementioned $(1-1/e)$\nand $1/e$ approximation ratios when $m = 0$ and $m = 1$, respectively. We\nfurther extend our results to consider a knapsack constraint. We show that a\nsampling-based policy achieves an approximation ratio of $(m+1)/10$ if the\nutility function is $m$-adaptive monotone and adaptive submodular. One\nimportant implication of our results is that even for a non-monotone utility\nfunction, we still can achieve an approximation ratio close to $(1-1/e)$ if\nthis function is ``close'' to a monotone function. This leads to improved\nperformance bounds for many machine learning applications whose utility\nfunctions are almost adaptive monotone.\n", "date": "2022-07-26", "categories": []} {"id": "2212.05192", "title": "Walkability Optimization: Formulations, Algorithms, and a Case Study of\n Toronto", "abstract": " The concept of walkable urban development has gained increased attention due\nto its public health, economic, and environmental sustainability benefits.\nUnfortunately, land zoning and historic under-investment have resulted in\nspatial inequality in walkability and social inequality among residents. We\ntackle the problem of Walkability Optimization through the lens of\ncombinatorial optimization. The task is to select locations in which additional\namenities (e.g., grocery stores, schools, restaurants) can be allocated to\nimprove resident access via walking while taking into account existing\namenities and providing multiple options (e.g., for restaurants). To this end,\nwe derive Mixed-Integer Linear Programming (MILP) and Constraint Programming\n(CP) models. Moreover, we show that the problem's objective function is\nsubmodular in special cases, which motivates an efficient greedy heuristic. We\nconduct a case study on 31 underserved neighborhoods in the City of Toronto,\nCanada. MILP finds the best solutions in most scenarios but does not scale well\nwith network size. The greedy algorithm scales well and finds near-optimal\nsolutions. Our empirical evaluation shows that neighbourhoods with low\nwalkability have a great potential for transformation into pedestrian-friendly\nneighbourhoods by strategically placing new amenities. Allocating 3 additional\ngrocery stores, schools, and restaurants can improve the \"WalkScore\" by more\nthan 50 points (on a scale of 100) for 4 neighbourhoods and reduce the walking\ndistances to amenities for 75% of all residential locations to 10 minutes for\nall amenity types. Our code and paper appendix are available at\nhttps://github.com/khalil-research/walkability.\n", "date": "2022-12-10", "categories": []} {"id": "2212.05412", "title": "A Hierarchical Temporal Planning-Based Approach for Dynamic Hoist\n Scheduling Problems", "abstract": " Hoist scheduling has become a bottleneck in electroplating industry\napplications with the development of autonomous devices. Although there are a\nfew approaches proposed to target at the challenging problem, they generally\ncannot scale to large-scale scheduling problems. In this paper, we formulate\nthe hoist scheduling problem as a new temporal planning problem in the form of\nadapted PDDL, and propose a novel hierarchical temporal planning approach to\nefficiently solve the scheduling problem. Additionally, we provide a collection\nof real-life benchmark instances that can be used to evaluate solution methods\nfor the problem. We exhibit that the proposed approach is able to efficiently\nfind solutions of high quality for large-scale real-life benchmark instances,\nwith comparison to state-of-the-art baselines.\n", "date": "2022-12-11", "categories": []} {"id": "2207.13513", "title": "Learning with Combinatorial Optimization Layers: a Probabilistic\n Approach", "abstract": " Combinatorial optimization (CO) layers in machine learning (ML) pipelines are\na powerful tool to tackle data-driven decision tasks, but they come with two\nmain challenges. First, the solution of a CO problem often behaves as a\npiecewise constant function of its objective parameters. Given that ML\npipelines are typically trained using stochastic gradient descent, the absence\nof slope information is very detrimental. Second, standard ML losses do not\nwork well in combinatorial settings. A growing body of research addresses these\nchallenges through diverse methods. Unfortunately, the lack of well-maintained\nimplementations slows down the adoption of CO layers.\n In this paper, building upon previous works, we introduce a probabilistic\nperspective on CO layers, which lends itself naturally to approximate\ndifferentiation and the construction of structured losses. We recover many\napproaches from the literature as special cases, and we also derive new ones.\nBased on this unifying perspective, we present InferOpt.jl, an open-source\nJulia package that 1) allows turning any CO oracle with a linear objective into\na differentiable layer, and 2) defines adequate losses to train pipelines\ncontaining such layers. Our library works with arbitrary optimization\nalgorithms, and it is fully compatible with Julia's ML ecosystem. We\ndemonstrate its abilities using a pathfinding problem on video game maps as\nguiding example, as well as three other applications from operations research.\n", "date": "2022-07-27", "categories": []} {"id": "2212.02228", "title": "First Competitive Ant Colony Scheme for the CARP", "abstract": " This paper addresses the Capacitated Arc Routing Problem (CARP) using an Ant\nColony Optimization scheme. Ant Colony schemes can compute solutions for medium\nscale instances of VRP. The proposed Ant Colony is dedicated to large-scale\ninstances of CARP with more than 140 nodes and 190 arcs to service. The Ant\nColony scheme is coupled with a local search procedure and provides high\nquality solutions. The benchmarks we carried out prove possible to obtain\nsolutions as profitable as CARPET ones can be obtained using such scheme when a\nsufficient number of iterations is devoted to the ants. It competes with the\nGenetic Algorithm of Lacomme et al. regarding solution quality but it is more\ntime consuming on large scale instances. The method has been intensively\nbenchmarked on the well-known instances of Eglese, DeArmon and the last ones of\nBelenguer and Benavent. This research report is a step forward CARP resolution\nby Ant Colony proving ant schemes can compete with Taboo search methods and\nGenetic Algorithms\n", "date": "2022-11-19", "categories": []} {"id": "2212.00639", "title": "Launchpad: Learning to Schedule Using Offline and Online RL Methods", "abstract": " Deep reinforcement learning algorithms have succeeded in several challenging\ndomains. Classic Online RL job schedulers can learn efficient scheduling\nstrategies but often takes thousands of timesteps to explore the environment\nand adapt from a randomly initialized DNN policy. Existing RL schedulers\noverlook the importance of learning from historical data and improving upon\ncustom heuristic policies. Offline reinforcement learning presents the prospect\nof policy optimization from pre-recorded datasets without online environment\ninteraction. Following the recent success of data-driven learning, we explore\ntwo RL methods: 1) Behaviour Cloning and 2) Offline RL, which aim to learn\npolicies from logged data without interacting with the environment. These\nmethods address the challenges concerning the cost of data collection and\nsafety, particularly pertinent to real-world applications of RL. Although the\ndata-driven RL methods generate good results, we show that the performance is\nhighly dependent on the quality of the historical datasets. Finally, we\ndemonstrate that by effectively incorporating prior expert demonstrations to\npre-train the agent, we short-circuit the random exploration phase to learn a\nreasonable policy with online training. We utilize Offline RL as a launchpad to\nlearn effective scheduling policies from prior experience collected using\nOracle or heuristic policies. Such a framework is effective for pre-training\nfrom historical datasets and well suited to continuous improvement with online\ndata collection.\n", "date": "2022-12-01", "categories": []} {"id": "2212.00121", "title": "Denoising Diffusion for Sampling SAT Solutions", "abstract": " Generating diverse solutions to the Boolean Satisfiability Problem (SAT) is a\nhard computational problem with practical applications for testing and\nfunctional verification of software and hardware designs. We explore the way to\ngenerate such solutions using Denoising Diffusion coupled with a Graph Neural\nNetwork to implement the denoising function. We find that the obtained accuracy\nis similar to the currently best purely neural method and the produced SAT\nsolutions are highly diverse, even if the system is trained with non-random\nsolutions from a standard solver.\n", "date": "2022-11-30", "categories": []} {"id": "2211.17193", "title": "Metaheuristic Approach to Solve Portfolio Selection Problem", "abstract": " In this paper, a heuristic method based on TabuSearch and TokenRing Search is\nbeing used in order to solve the Portfolio Optimization Problem. The seminal\nmean-variance model of Markowitz is being considered with the addition of\ncardinality and quantity constraints to better capture the dynamics of the\ntrading procedure, the model becomes an NP-hard problem that can not be solved\nusing an exact method. The combination of three different neighborhood\nrelations is being explored with Tabu Search. In addition, a new constructive\nmethod for the initial solution is proposed. Finally, I show how the proposed\ntechniques perform on public benchmarks\n", "date": "2022-11-10", "categories": []} {"id": "2211.17199", "title": "Resource Sharing Through Multi-Round Matchings", "abstract": " Applications such as employees sharing office spaces over a workweek can be\nmodeled as problems where agents are matched to resources over multiple rounds.\nAgents' requirements limit the set of compatible resources and the rounds in\nwhich they want to be matched. Viewing such an application as a multi-round\nmatching problem on a bipartite compatibility graph between agents and\nresources, we show that a solution (i.e., a set of matchings, with one matching\nper round) can be found efficiently if one exists. To cope with situations\nwhere a solution does not exist, we consider two extensions. In the first\nextension, a benefit function is defined for each agent and the objective is to\nfind a multi-round matching to maximize the total benefit. For a general class\nof benefit functions satisfying certain properties (including diminishing\nreturns), we show that this multi-round matching problem is efficiently\nsolvable. This class includes utilitarian and Rawlsian welfare functions. For\nanother benefit function, we show that the maximization problem is NP-hard. In\nthe second extension, the objective is to generate advice to each agent (i.e.,\na subset of requirements to be relaxed) subject to a budget constraint so that\nthe agent can be matched. We show that this budget-constrained advice\ngeneration problem is NP-hard. For this problem, we develop an integer linear\nprogramming formulation as well as a heuristic based on local search. We\nexperimentally evaluate our algorithms on synthetic networks and apply them to\ntwo real-world situations: shared office spaces and matching courses to\nclassrooms.\n", "date": "2022-11-30", "categories": []} {"id": "2211.16011", "title": "Incorporating Multi-armed Bandit with Local Search for MaxSAT", "abstract": " Partial MaxSAT (PMS) and Weighted PMS (WPMS) are two practical\ngeneralizations of the MaxSAT problem. In this paper, we propose a local search\nalgorithm for these problems, called BandHS, which applies two multi-armed\nbandits to guide the search directions when escaping local optima. One bandit\nis combined with all the soft clauses to help the algorithm select to satisfy\nappropriate soft clauses, and the other bandit with all the literals in hard\nclauses to help the algorithm select appropriate literals to satisfy the hard\nclauses. These two bandits can improve the algorithm's search ability in both\nfeasible and infeasible solution spaces. We further propose an initialization\nmethod for (W)PMS that prioritizes both unit and binary clauses when producing\nthe initial solutions. Extensive experiments demonstrate the excellent\nperformance and generalization capability of our proposed methods, that greatly\nboost the state-of-the-art local search algorithm, SATLike3.0, and the\nstate-of-the-art SAT-based incomplete solver, NuWLS-c.\n", "date": "2022-11-29", "categories": []} {"id": "2211.14405", "title": "Learning Branching Heuristics from Graph Neural Networks", "abstract": " Backtracking has been widely used for solving problems in artificial\nintelligence (AI), including constraint satisfaction problems and combinatorial\noptimization problems. Good branching heuristics can efficiently improve the\nperformance of backtracking by helping prune the search space and leading the\nsearch to the most promising direction. In this paper, we first propose a new\ngraph neural network (GNN) model designed using the probabilistic method. From\nthe GNN model, we introduce an approach to learn a branching heuristic for\ncombinatorial optimization problems. In particular, our GNN model learns\nappropriate probability distributions on vertices in given graphs from which\nthe branching heuristic is extracted and used in a backtracking search. Our\nexperimental results for the (minimum) dominating-clique problem show that this\nlearned branching heuristic performs better than the minimum-remaining-values\nheuristic in terms of the number of branches of the whole search tree. Our\napproach introduces a new way of applying GNNs towards enhancing the classical\nbacktracking algorithm used in AI.\n", "date": "2022-11-26", "categories": []} {"id": "2211.13432", "title": "LaCAM: Search-Based Algorithm for Quick Multi-Agent Pathfinding", "abstract": " We propose a novel complete algorithm for multi-agent pathfinding (MAPF)\ncalled lazy constraints addition search for MAPF (LaCAM). MAPF is a problem of\nfinding collision-free paths for multiple agents on graphs and is the\nfoundation of multi-robot coordination. LaCAM uses a two-level search to find\nsolutions quickly, even with hundreds of agents or more. At the low-level, it\nsearches constraints about agents' locations. At the high-level, it searches a\nsequence of all agents' locations, following the constraints specified by the\nlow-level. Our exhaustive experiments reveal that LaCAM is comparable to or\noutperforms state-of-the-art sub-optimal MAPF algorithms in a variety of\nscenarios, regarding success rate, planning time, and solution quality of\nsum-of-costs.\n", "date": "2022-11-24", "categories": []} {"id": "2211.13922", "title": "Combining Constructive and Perturbative Deep Learning Algorithms for the\n Capacitated Vehicle Routing Problem", "abstract": " The Capacitated Vehicle Routing Problem is a well-known NP-hard problem that\nposes the challenge of finding the optimal route of a vehicle delivering\nproducts to multiple locations. Recently, new efforts have emerged to create\nconstructive and perturbative heuristics to tackle this problem using Deep\nLearning. In this paper, we join these efforts to develop the Combined Deep\nConstructor and Perturbator, which combines two powerful constructive and\nperturbative Deep Learning-based heuristics, using attention mechanisms at\ntheir core. Furthermore, we improve the Attention Model-Dynamic for the\nCapacitated Vehicle Routing Problem by proposing a memory-efficient algorithm\nthat reduces its memory complexity by a factor of the number of nodes. Our\nmethod shows promising results. It demonstrates a cost improvement in common\ndatasets when compared against other multiple Deep Learning methods. It also\nobtains close results to the state-of-the art heuristics from the Operations\nResearch field. Additionally, the proposed memory efficient algorithm for the\nAttention Model-Dynamic model enables its use in problem instances with more\nthan 100 nodes.\n", "date": "2022-11-25", "categories": []} {"id": "2211.12885", "title": "Cost Splitting for Multi-Objective Conflict-Based Search", "abstract": " The Multi-Objective Multi-Agent Path Finding (MO-MAPF) problem is the problem\nof finding the Pareto-optimal frontier of collision-free paths for a team of\nagents while minimizing multiple cost metrics. Examples of such cost metrics\ninclude arrival times, travel distances, and energy consumption.In this paper,\nwe focus on the Multi-Objective Conflict-Based Search (MO-CBS) algorithm, a\nstate-of-the-art MO-MAPF algorithm. We show that the standard splitting\nstrategy used by MO-CBS can lead to duplicate search nodes and hence can\nduplicate the search effort that MO-CBS needs to make. To address this issue,\nwe propose two new splitting strategies for MO-CBS, namely cost splitting and\ndisjoint cost splitting. Our theoretical results show that, when combined with\neither of these two new splitting strategies, MO-CBS maintains its completeness\nand optimality guarantees. Our experimental results show that disjoint cost\nsplitting, our best splitting strategy, speeds up MO-CBS by up to two orders of\nmagnitude and substantially improves its success rates in various settings.\n", "date": "2022-11-23", "categories": []} {"id": "2211.11564", "title": "Adaptive Constraint Partition based Optimization Framework for\n Large-scale Integer Linear Programming(Student Abstract)", "abstract": " Integer programming problems (IPs) are challenging to be solved efficiently\ndue to the NP-hardness, especially for large-scale IPs. To solve this type of\nIPs, Large neighborhood search (LNS) uses an initial feasible solution and\niteratively improves it by searching a large neighborhood around the current\nsolution. However, LNS easily steps into local optima and ignores the\ncorrelation between variables to be optimized, leading to compromised\nperformance. This paper presents a general adaptive constraint partition-based\noptimization framework (ACP) for large-scale IPs that can efficiently use any\nexisting optimization solver as a subroutine. Specifically, ACP first randomly\npartitions the constraints into blocks, where the number of blocks is\nadaptively adjusted to avoid local optima. Then, ACP uses a subroutine solver\nto optimize the decision variables in a randomly selected block of constraints\nto enhance the variable correlation. ACP is compared with LNS framework with\ndifferent subroutine solvers on four IPs and a real-world IP. The experimental\nresults demonstrate that in specified wall-clock time ACP shows better\nperformance than SCIP and Gurobi.\n", "date": "2022-11-18", "categories": []} {"id": "2207.06190", "title": "Simulation-guided Beam Search for Neural Combinatorial Optimization", "abstract": " Neural approaches for combinatorial optimization (CO) equip a learning\nmechanism to discover powerful heuristics for solving complex real-world\nproblems. While neural approaches capable of high-quality solutions in a single\nshot are emerging, state-of-the-art approaches are often unable to take full\nadvantage of the solving time available to them. In contrast, hand-crafted\nheuristics perform highly effective search well and exploit the computation\ntime given to them, but contain heuristics that are difficult to adapt to a\ndataset being solved. With the goal of providing a powerful search procedure to\nneural CO approaches, we propose simulation-guided beam search (SGBS), which\nexamines candidate solutions within a fixed-width tree search that both a\nneural net-learned policy and a simulation (rollout) identify as promising. We\nfurther hybridize SGBS with efficient active search (EAS), where SGBS enhances\nthe quality of solutions backpropagated in EAS, and EAS improves the quality of\nthe policy used in SGBS. We evaluate our methods on well-known CO benchmarks\nand show that SGBS significantly improves the quality of the solutions found\nunder reasonable runtime assumptions.\n", "date": "2022-07-13", "categories": []} {"id": "2211.09665", "title": "Features for the 0-1 knapsack problem based on inclusionwise maximal\n solutions", "abstract": " Decades of research on the 0-1 knapsack problem led to very efficient\nalgorithms that are able to quickly solve large problem instances to\noptimality. This prompted researchers to also investigate whether relatively\nsmall problem instances exist that are hard for existing solvers and\ninvestigate which features characterize their hardness. Previously the authors\nproposed a new class of hard 0-1 knapsack problem instances and demonstrated\nthat the properties of so-called inclusionwise maximal solutions (IMSs) can be\nimportant hardness indicators for this class. In the current paper, we\nformulate several new computationally challenging problems related to the IMSs\nof arbitrary 0-1 knapsack problem instances. Based on generalizations of\nprevious work and new structural results about IMSs, we formulate polynomial\nand pseudopolynomial time algorithms for solving these problems. From this we\nderive a set of 14 computationally expensive features, which we calculate for\ntwo large datasets on a supercomputer in approximately 540 CPU-hours. We show\nthat the proposed features contain important information related to the\nempirical hardness of a problem instance that was missing in earlier features\nfrom the literature by training machine learning models that can accurately\npredict the empirical hardness of a wide variety of 0-1 knapsack problem\ninstances. Using the instance space analysis methodology, we also show that\nhard 0-1 knapsack problem instances are clustered together around a relatively\ndense region of the instance space and several features behave differently in\nthe easy and hard parts of the instance space.\n", "date": "2022-11-16", "categories": []} {"id": "2206.04423", "title": "Learning to generalize Dispatching rules on the Job Shop Scheduling", "abstract": " This paper introduces a Reinforcement Learning approach to better generalize\nheuristic dispatching rules on the Job-shop Scheduling Problem (JSP). Current\nmodels on the JSP do not focus on generalization, although, as we show in this\nwork, this is key to learning better heuristics on the problem. A well-known\ntechnique to improve generalization is to learn on increasingly complex\ninstances using Curriculum Learning (CL). However, as many works in the\nliterature indicate, this technique might suffer from catastrophic forgetting\nwhen transferring the learned skills between different problem sizes. To\naddress this issue, we introduce a novel Adversarial Curriculum Learning (ACL)\nstrategy, which dynamically adjusts the difficulty level during the learning\nprocess to revisit the worst-performing instances. This work also presents a\ndeep learning model to solve the JSP, which is equivariant w.r.t. the job\ndefinition and size-agnostic. Conducted experiments on Taillard's and\nDemirkol's instances show that the presented approach significantly improves\nthe current state-of-the-art models on the JSP. It reduces the average\noptimality gap from 19.35\\% to 10.46\\% on Taillard's instances and from 38.43\\%\nto 18.85\\% on Demirkol's instances. Our implementation is available online.\n", "date": "2022-06-09", "categories": []} {"id": "2211.07000", "title": "Online Correlation Clustering for Dynamic Complete Signed Graphs", "abstract": " In the correlation clustering problem for complete signed graphs, the input\nis a complete signed graph with edges weighted as $+1$ (denote recommendation\nto put this pair in the same cluster) or $-1$ (recommending to put this pair of\nvertices in separate clusters) and the target is to cluster the set of vertices\nsuch that the number of disagreements with these recommendations is minimized.\n In this paper, we consider the problem of correlation clustering for dynamic\ncomplete signed graphs where (1) a vertex can be added or deleted, and (2) the\nsign of an edge can be flipped. In the proposed online scheme, the offline\napproximation algorithm in [CALM+21] for correlation clustering is used. Up to\nthe author's knowledge, this is the first online algorithm for dynamic graphs\nwhich allows a full set of graph editing operations.\n The proposed approach is rigorously analyzed and compared with a baseline\nmethod, which runs the original offline algorithm on each time step. Our\nresults show that the dynamic operations have local effects on the neighboring\nvertices and we employ this locality to reduce the dependency of the running\ntime in the Baseline to the summation of the degree of all vertices in $G_t$,\nthe graph after applying the graph edit operation at time step $t$, to the\nsummation of the degree of the changing vertices (e.g. two endpoints of an\nedge) and the number of clusters in the previous time step. Moreover, the\nrequired working memory is reduced to the square of the summation of the degree\nof the modified edge endpoints rather than the total number of vertices in the\ngraph.\n", "date": "2022-11-13", "categories": []} {"id": "2211.05297", "title": "Coordinating CAV Swarms at Intersections with a Deep Learning Model", "abstract": " Connected and automated vehicles (CAVs) are viewed as a special kind of\nrobots that have the potential to significantly improve the safety and\nefficiency of traffic. In contrast to many swarm robotics studies that are\ndemonstrated in labs by employing a small number of robots, CAV studies aims to\nachieve cooperative driving of unceasing robot swarm flows. However, how to get\nthe optimal passing order of such robot swarm flows even for a signal-free\nintersection is an NP-hard problem (specifically, enumerating based algorithm\ntakes days to find the optimal solution to a 20-CAV scenario). Here, we\nintroduce a novel cooperative driving algorithm (AlphaOrder) that combines\noffline deep learning and online tree searching to find a near-optimal passing\norder in real-time. AlphaOrder builds a pointer network model from solved\nscenarios and generates near-optimal passing orders instantaneously for new\nscenarios. Furthermore, our approach provides a general approach to managing\npreemptive resource sharing between swarm robotics (e.g., scheduling multiple\nautomated guided vehicles (AGVs) and unmanned aerial vehicles (UAVs) at\nconflicting areas\n", "date": "2022-11-10", "categories": []} {"id": "2211.03880", "title": "NSNet: A General Neural Probabilistic Framework for Satisfiability\n Problems", "abstract": " We present the Neural Satisfiability Network (NSNet), a general neural\nframework that models satisfiability problems as probabilistic inference and\nmeanwhile exhibits proper explainability. Inspired by the Belief Propagation\n(BP), NSNet uses a novel graph neural network (GNN) to parameterize BP in the\nlatent space, where its hidden representations maintain the same probabilistic\ninterpretation as BP. NSNet can be flexibly configured to solve both SAT and\n#SAT problems by applying different learning objectives. For SAT, instead of\ndirectly predicting a satisfying assignment, NSNet performs marginal inference\namong all satisfying solutions, which we empirically find is more feasible for\nneural networks to learn. With the estimated marginals, a satisfying assignment\ncan be efficiently generated by rounding and executing a stochastic local\nsearch. For #SAT, NSNet performs approximate model counting by learning the\nBethe approximation of the partition function. Our evaluations show that NSNet\nachieves competitive results in terms of inference accuracy and time efficiency\non multiple SAT and #SAT datasets.\n", "date": "2022-11-07", "categories": []} {"id": "2210.16963", "title": "Learning Heuristics for the Maximum Clique Enumeration Problem Using Low\n Dimensional Representations", "abstract": " Approximate solutions to various NP-hard combinatorial optimization problems\nhave been found by learned heuristics using complex learning models. In\nparticular, vertex (node) classification in graphs has been a helpful method\ntowards finding the decision boundary to distinguish vertices in an optimal set\nfrom the rest. By following this approach, we use a learning framework for a\npruning process of the input graph towards reducing the runtime of the maximum\nclique enumeration problem. We extensively study the role of using different\nvertex representations on the performance of this heuristic method, using graph\nembedding algorithms, such as Node2vec and DeepWalk, and representations using\nhigher-order graph features comprising local subgraph counts. Our results show\nthat Node2Vec and DeepWalk are promising embedding methods in representing\nnodes towards classification purposes. We observe that using local graph\nfeatures in the classification process produce more accurate results when\ncombined with a feature elimination process. Finally, we provide tests on\nrandom graphs to show the robustness and scalability of our method.\n", "date": "2022-10-30", "categories": []} {"id": "2210.15220", "title": "End-to-End Pareto Set Prediction with Graph Neural Networks for\n Multi-objective Facility Location", "abstract": " The facility location problems (FLPs) are a typical class of NP-hard\ncombinatorial optimization problems, which are widely seen in the supply chain\nand logistics. Many mathematical and heuristic algorithms have been developed\nfor optimizing the FLP. In addition to the transportation cost, there are\nusually multiple conflicting objectives in realistic applications. It is\ntherefore desirable to design algorithms that find a set of Pareto solutions\nefficiently without enormous search cost. In this paper, we consider the\nmulti-objective facility location problem (MO-FLP) that simultaneously\nminimizes the overall cost and maximizes the system reliability. We develop a\nlearning-based approach to predicting the distribution probability of the\nentire Pareto set for a given problem. To this end, the MO-FLP is modeled as a\nbipartite graph optimization problem and two graph neural networks are\nconstructed to learn the implicit graph representation on nodes and edges. The\nnetwork outputs are then converted into the probability distribution of the\nPareto set, from which a set of non-dominated solutions can be sampled\nnon-autoregressively. Experimental results on MO-FLP instances of different\nscales show that the proposed approach achieves a comparable performance to a\nwidely used multi-objective evolutionary algorithm in terms of the solution\nquality while significantly reducing the computational cost for search.\n", "date": "2022-10-27", "categories": []} {"id": "2210.04123", "title": "DIMES: A Differentiable Meta Solver for Combinatorial Optimization\n Problems", "abstract": " Recently, deep reinforcement learning (DRL) models have shown promising\nresults in solving NP-hard Combinatorial Optimization (CO) problems. However,\nmost DRL solvers can only scale to a few hundreds of nodes for combinatorial\noptimization problems on graphs, such as the Traveling Salesman Problem (TSP).\nThis paper addresses the scalability challenge in large-scale combinatorial\noptimization by proposing a novel approach, namely, DIMES. Unlike previous DRL\nmethods which suffer from costly autoregressive decoding or iterative\nrefinements of discrete solutions, DIMES introduces a compact continuous space\nfor parameterizing the underlying distribution of candidate solutions. Such a\ncontinuous space allows stable REINFORCE-based training and fine-tuning via\nmassively parallel sampling. We further propose a meta-learning framework to\nenable the effective initialization of model parameters in the fine-tuning\nstage. Extensive experiments show that DIMES outperforms recent DRL-based\nmethods on large benchmark datasets for Traveling Salesman Problems and Maximal\nIndependent Set problems.\n", "date": "2022-10-08", "categories": []} {"id": "2207.05984", "title": "Unsupervised Learning for Combinatorial Optimization with Principled\n Objective Relaxation", "abstract": " Using machine learning to solve combinatorial optimization (CO) problems is\nchallenging, especially when the data is unlabeled. This work proposes an\nunsupervised learning framework for CO problems. Our framework follows a\nstandard relaxation-plus-rounding approach and adopts neural networks to\nparameterize the relaxed solutions so that simple back-propagation can train\nthe model end-to-end. Our key contribution is the observation that if the\nrelaxed objective satisfies entry-wise concavity, a low optimization loss\nguarantees the quality of the final integral solutions. This observation\nsignificantly broadens the applicability of the previous framework inspired by\nErdos' probabilistic method. In particular, this observation can guide the\ndesign of objective models in applications where the objectives are not given\nexplicitly while requiring being modeled in prior. We evaluate our framework by\nsolving a synthetic graph optimization problem, and two real-world applications\nincluding resource allocation in circuit design and approximate computing. Our\nframework largely outperforms the baselines based on na\\\"{i}ve relaxation,\nreinforcement learning, and Gumbel-softmax tricks.\n", "date": "2022-07-13", "categories": []} {"id": "2210.12075", "title": "Neural Networks for Local Search and Crossover in Vehicle Routing: A\n Possible Overkill?", "abstract": " Extensive research has been conducted, over recent years, on various ways of\nenhancing heuristic search for combinatorial optimization problems with machine\nlearning algorithms. In this study, we investigate the use of predictions from\ngraph neural networks (GNNs) in the form of heatmaps to improve the Hybrid\nGenetic Search (HGS), a state-of-the-art algorithm for the Capacitated Vehicle\nRouting Problem (CVRP). The crossover and local-search components of HGS are\ninstrumental in finding improved solutions, yet these components essentially\nrely on simple greedy or random choices. It seems intuitive to attempt to\nincorporate additional knowledge at these levels. Throughout a vast\nexperimental campaign on more than 10,000 problem instances, we show that\nexploiting more sophisticated strategies using measures of node relatedness\n(heatmaps, or simply distance) within these algorithmic components can\nsignificantly enhance performance. However, contrary to initial expectations,\nwe also observed that heatmaps did not present significant advantages over\nsimpler distance measures for these purposes. Therefore, we faced a common --\nthough rarely documented -- situation of overkill: GNNs can indeed improve\nperformance on an important optimization task, but an ablation analysis\ndemonstrated that simpler alternatives perform equally well.\n", "date": "2022-09-09", "categories": []} {"id": "2210.09927", "title": "Research of an optimization model for servicing a network of ATMs and\n information payment terminals", "abstract": " The steadily high demand for cash contributes to the expansion of the network\nof Bank payment terminals. To optimize the amount of cash in payment terminals,\nit is necessary to minimize the cost of servicing them and ensure that there\nare no excess funds in the network. The purpose of this work is to create a\ncash management system in the network of payment terminals. The article\ndiscusses the solution to the problem of determining the optimal amount of\nfunds to be loaded into the terminals, and the effective frequency of\ncollection, which allows to get additional income by investing the released\nfunds. The paper presents the results of predicting daily cash withdrawals at\nATMs using a triple exponential smoothing model, a recurrent neural network\nwith long short-term memory, and a model of singular spectrum analysis. These\nforecasting models allowed us to obtain a sufficient level of correct forecasts\nwith good accuracy and completeness. The results of forecasting cash\nwithdrawals were used to build a discrete optimal control model, which was used\nto develop an optimal schedule for adding funds to the payment terminal. It is\nproved that the efficiency and reliability of the proposed model is higher than\nthat of the classical Baumol-Tobin inventory management model: when tested on\nthe time series of three ATMs, the discrete optimal control model did not allow\nexhaustion of funds and allowed to earn on average 30% more than the classical\nmodel.\n", "date": "2022-10-18", "categories": []} {"id": "2210.05236", "title": "Planning Assembly Sequence with Graph Transformer", "abstract": " Assembly sequence planning (ASP) is the essential process for modern\nmanufacturing, proven to be NP-complete thus its effective and efficient\nsolution has been a challenge for researchers in the field. In this paper, we\npresent a graph-transformer based framework for the ASP problem which is\ntrained and demonstrated on a self-collected ASP database. The ASP database\ncontains a self-collected set of LEGO models. The LEGO model is abstracted to a\nheterogeneous graph structure after a thorough analysis of the original\nstructure and feature extraction. The ground truth assembly sequence is first\ngenerated by brute-force search and then adjusted manually to in line with\nhuman rational habits. Based on this self-collected ASP dataset, we propose a\nheterogeneous graph-transformer framework to learn the latent rules for\nassembly planning. We evaluated the proposed framework in a series of\nexperiment. The results show that the similarity of the predicted and ground\ntruth sequences can reach 0.44, a medium correlation measured by Kendall's\n$\\tau$. Meanwhile, we compared the different effects of node features and edge\nfeatures and generated a feasible and reasonable assembly sequence as a\nbenchmark for further research. Our data set and code is available on\nhttps://github.com/AIR-DISCOVER/ICRA\\_ASP.\n", "date": "2022-10-11", "categories": []} {"id": "2210.05906", "title": "Travel the Same Path: A Novel TSP Solving Strategy", "abstract": " In this paper, we provide a novel strategy for solving Traveling Salesman\nProblem, which is a famous combinatorial optimization problem studied intensely\nin the TCS community. In particular, we consider the imitation learning\nframework, which helps a deterministic algorithm making good choices whenever\nit needs to, resulting in a speed up while maintaining the exactness of the\nsolution without suffering from the unpredictability and a potential large\ndeviation.\n Furthermore, we demonstrate a strong generalization ability of a graph neural\nnetwork trained under the imitation learning framework. Specifically, the model\nis capable of solving a large instance of TSP faster than the baseline while\nhas only seen small TSP instances when training.\n", "date": "2022-10-12", "categories": []} {"id": "2207.05899", "title": "Neural Topological Ordering for Computation Graphs", "abstract": " Recent works on machine learning for combinatorial optimization have shown\nthat learning based approaches can outperform heuristic methods in terms of\nspeed and performance. In this paper, we consider the problem of finding an\noptimal topological order on a directed acyclic graph with focus on the memory\nminimization problem which arises in compilers. We propose an end-to-end\nmachine learning based approach for topological ordering using an\nencoder-decoder framework. Our encoder is a novel attention based graph neural\nnetwork architecture called \\emph{Topoformer} which uses different topological\ntransforms of a DAG for message passing. The node embeddings produced by the\nencoder are converted into node priorities which are used by the decoder to\ngenerate a probability distribution over topological orders. We train our model\non a dataset of synthetically generated graphs called layered graphs. We show\nthat our model outperforms, or is on-par, with several topological ordering\nbaselines while being significantly faster on synthetic graphs with up to 2k\nnodes. We also train and test our model on a set of real-world computation\ngraphs, showing performance improvements.\n", "date": "2022-07-13", "categories": []} {"id": "2210.03674", "title": "Reinforcement Learning Approach for Multi-Agent Flexible Scheduling\n Problems", "abstract": " Scheduling plays an important role in automated production. Its impact can be\nfound in various fields such as the manufacturing industry, the service\nindustry and the technology industry. A scheduling problem (NP-hard) is a task\nof finding a sequence of job assignments on a given set of machines with the\ngoal of optimizing the objective defined. Methods such as Operation Research,\nDispatching Rules, and Combinatorial Optimization have been applied to\nscheduling problems but no solution guarantees to find the optimal solution.\nThe recent development of Reinforcement Learning has shown success in\nsequential decision-making problems. This research presents a Reinforcement\nLearning approach for scheduling problems. In particular, this study delivers\nan OpenAI gym environment with search-space reduction for Job Shop Scheduling\nProblems and provides a heuristic-guided Q-Learning solution with\nstate-of-the-art performance for Multi-agent Flexible Job Shop Problems.\n", "date": "2022-10-07", "categories": []} {"id": "2206.00787", "title": "On the Generalization of Neural Combinatorial Optimization Heuristics", "abstract": " Neural Combinatorial Optimization approaches have recently leveraged the\nexpressiveness and flexibility of deep neural networks to learn efficient\nheuristics for hard Combinatorial Optimization (CO) problems. However, most of\nthe current methods lack generalization: for a given CO problem, heuristics\nwhich are trained on instances with certain characteristics underperform when\ntested on instances with different characteristics. While some previous works\nhave focused on varying the training instances properties, we postulate that a\none-size-fit-all model is out of reach. Instead, we formalize solving a CO\nproblem over a given instance distribution as a separate learning task and\ninvestigate meta-learning techniques to learn a model on a variety of tasks, in\norder to optimize its capacity to adapt to new tasks. Through extensive\nexperiments, on two CO problems, using both synthetic and realistic instances,\nwe show that our proposed meta-learning approach significantly improves the\ngeneralization of two state-of-the-art models.\n", "date": "2022-06-01", "categories": []} {"id": "2209.15457", "title": "Scheduling for Urban Air Mobility using Safe Learning", "abstract": " This work considers the scheduling problem for Urban Air Mobility (UAM)\nvehicles travelling between origin-destination pairs with both hard and soft\ntrip deadlines. Each route is described by a discrete probability distribution\nover trip completion times (or delay) and over inter-arrival times of requests\n(or demand) for the route along with a fixed hard or soft deadline. Soft\ndeadlines carry a cost that is incurred when the deadline is missed. An online,\nsafe scheduler is developed that ensures that hard deadlines are never missed,\nand that average cost of missing soft deadlines is minimized. The system is\nmodelled as a Markov Decision Process (MDP) and safe model-based learning is\nused to find the probabilistic distributions over route delays and demand.\nMonte Carlo Tree Search (MCTS) Earliest Deadline First (EDF) is used to safely\nexplore the learned models in an online fashion and develop a near-optimal\nnon-preemptive scheduling policy. These results are compared with Value\nIteration (VI) and MCTS (Random) scheduling solutions.\n", "date": "2022-09-28", "categories": []} {"id": "2209.13077", "title": "Accelerating the Genetic Algorithm for Large-scale Traveling Salesman\n Problems by Cooperative Coevolutionary Pointer Network with Reinforcement\n Learning", "abstract": " In this paper, we propose a two-stage optimization strategy for solving the\nLarge-scale Traveling Salesman Problems (LSTSPs) named CCPNRL-GA. First, we\nhypothesize that the participation of a well-performed individual as an elite\ncan accelerate the convergence of optimization. Based on this hypothesis, in\nthe first stage, we cluster the cities and decompose the LSTSPs into multiple\nsubcomponents, and each subcomponent is optimized with a reusable Pointer\nNetwork (PtrNet). After subcomponents optimization, we combine all sub-tours to\nform a valid solution, this solution joins the second stage of optimization\nwith GA. We validate the performance of our proposal on 10 LSTSPs and compare\nit with traditional EAs. Experimental results show that the participation of an\nelite individual can greatly accelerate the optimization of LSTSPs, and our\nproposal has broad prospects for dealing with LSTSPs.\n", "date": "2022-09-27", "categories": []} {"id": "2209.12000", "title": "Deep Attentive Belief Propagation: Integrating Reasoning and Learning\n for Solving Constraint Optimization Problems", "abstract": " Belief Propagation (BP) is an important message-passing algorithm for various\nreasoning tasks over graphical models, including solving the Constraint\nOptimization Problems (COPs). It has been shown that BP can achieve\nstate-of-the-art performance on various benchmarks by mixing old and new\nmessages before sending the new one, i.e., damping. However, existing methods\nof tuning a static damping factor for BP not only are laborious but also harm\ntheir performance. Moreover, existing BP algorithms treat each variable node's\nneighbors equally when composing a new message, which also limits their\nexploration ability. To address these issues, we seamlessly integrate BP, Gated\nRecurrent Units (GRUs), and Graph Attention Networks (GATs) within the\nmessage-passing framework to reason about dynamic weights and damping factors\nfor composing new BP messages. Our model, Deep Attentive Belief Propagation\n(DABP), takes the factor graph and the BP messages in each iteration as the\ninput and infers the optimal weights and damping factors through GRUs and GATs,\nfollowed by a multi-head attention layer. Furthermore, unlike existing\nneural-based BP variants, we propose a novel self-supervised learning algorithm\nfor DABP with a smoothed solution cost, which does not require expensive\ntraining labels and also avoids the common out-of-distribution issue through\nefficient online learning. Extensive experiments show that our model\nsignificantly outperforms state-of-the-art baselines.\n", "date": "2022-09-24", "categories": []} {"id": "2209.09983", "title": "Deep-Steiner: Learning to Solve the Euclidean Steiner Tree Problem", "abstract": " The Euclidean Steiner tree problem seeks the min-cost network to connect a\ncollection of target locations, and it underlies many applications of wireless\nnetworks. In this paper, we present a study on solving the Euclidean Steiner\ntree problem using reinforcement learning enhanced by graph representation\nlearning. Different from the commonly studied connectivity problems like\ntravelling salesman problem or vehicle routing problem where the search space\nis finite, the Euclidean Steiner tree problem requires to search over the\nentire Euclidean space, thereby making the existing methods not applicable. In\nthis paper, we design discretization methods by leveraging the unique\ncharacteristics of the Steiner tree, and propose new training schemes for\nhandling the dynamic Steiner points emerging during the incremental\nconstruction. Our design is examined through a sanity check using experiments\non a collection of datasets, with encouraging results demonstrating the utility\nof our method as an alternative to classic combinatorial methods.\n", "date": "2022-09-20", "categories": []} {"id": "2209.09612", "title": "Analysis Of The Anytime MAPF Solvers Based On The Combination Of\n Conflict-Based Search (CBS) and Focal Search (FS)", "abstract": " Conflict-Based Search (CBS) is a widely used algorithm for solving\nmulti-agent pathfinding (MAPF) problems optimally. The core idea of CBS is to\nrun hierarchical search, when, on the high level the tree of solutions\ncandidates is explored, and on the low-level an individual planning for a\nspecific agent (subject to certain constraints) is carried out. To trade-off\noptimality for running time different variants of bounded sub-optimal CBS were\ndesigned, which alter both high- and low-level search routines of CBS.\nMoreover, anytime variant of CBS does exist that applies Focal Search (FS) to\nthe high-level of CBS - Anytime BCBS. However, no comprehensive analysis of how\nwell this algorithm performs compared to the naive one, when we simply\nre-invoke CBS with the decreased sub-optimality bound, was present. This work\naims at filling this gap. Moreover, we present and evaluate another anytime\nversion of CBS that uses FS on both levels of CBS. Empirically, we show that\nits behavior is principally different from the one demonstrated by Anytime\nBCBS. Finally, we compare both algorithms head-to-head and show that using\nFocal Search on both levels of CBS can be beneficial in a wide range of setups.\n", "date": "2022-09-20", "categories": []} {"id": "2207.01443", "title": "Solving the Traveling Salesperson Problem with Precedence Constraints by\n Deep Reinforcement Learning", "abstract": " This work presents solutions to the Traveling Salesperson Problem with\nprecedence constraints (TSPPC) using Deep Reinforcement Learning (DRL) by\nadapting recent approaches that work well for regular TSPs. Common to these\napproaches is the use of graph models based on multi-head attention (MHA)\nlayers. One idea for solving the pickup and delivery problem (PDP) is using\nheterogeneous attentions to embed the different possible roles each node can\ntake. In this work, we generalize this concept of heterogeneous attentions to\nthe TSPPC. Furthermore, we adapt recent ideas to sparsify attentions for better\nscalability. Overall, we contribute to the research community through the\napplication and evaluation of recent DRL methods in solving the TSPPC.\n", "date": "2022-07-04", "categories": []} {"id": "2207.03244", "title": "Learning the Quality of Machine Permutations in Job Shop Scheduling", "abstract": " In recent years, the power demonstrated by Machine Learning (ML) has\nincreasingly attracted the interest of the optimization community that is\nstarting to leverage ML for enhancing and automating the design of algorithms.\nOne combinatorial optimization problem recently tackled with ML is the Job Shop\nscheduling Problem (JSP). Most of the works on the JSP using ML focus on Deep\nReinforcement Learning (DRL), and only a few of them leverage supervised\nlearning techniques. The recurrent reasons for avoiding supervised learning\nseem to be the difficulty in casting the right learning task, i.e., what is\nmeaningful to predict, and how to obtain labels. Therefore, we first propose a\nnovel supervised learning task that aims at predicting the quality of machine\npermutations. Then, we design an original methodology to estimate this quality,\nand we use these estimations to create an accurate sequential deep learning\nmodel (binary accuracy above 95%). Finally, we empirically demonstrate the\nvalue of predicting the quality of machine permutations by enhancing the\nperformance of a simple Tabu Search algorithm inspired by the works in the\nliterature.\n", "date": "2022-07-07", "categories": []} {"id": "2209.04698", "title": "Structured Q-learning For Antibody Design", "abstract": " Optimizing combinatorial structures is core to many real-world problems, such\nas those encountered in life sciences. For example, one of the crucial steps\ninvolved in antibody design is to find an arrangement of amino acids in a\nprotein sequence that improves its binding with a pathogen. Combinatorial\noptimization of antibodies is difficult due to extremely large search spaces\nand non-linear objectives. Even for modest antibody design problems, where\nproteins have a sequence length of eleven, we are faced with searching over\n2.05 x 10^14 structures. Applying traditional Reinforcement Learning algorithms\nsuch as Q-learning to combinatorial optimization results in poor performance.\nWe propose Structured Q-learning (SQL), an extension of Q-learning that\nincorporates structural priors for combinatorial optimization. Using a\nmolecular docking simulator, we demonstrate that SQL finds high binding energy\nsequences and performs favourably against baselines on eight challenging\nantibody design tasks, including designing antibodies for SARS-COV.\n", "date": "2022-09-10", "categories": []} {"id": "2209.06094", "title": "Learning to Solve Multiple-TSP with Time Window and Rejections via Deep\n Reinforcement Learning", "abstract": " We propose a manager-worker framework based on deep reinforcement learning to\ntackle a hard yet nontrivial variant of Travelling Salesman Problem (TSP),\n\\ie~multiple-vehicle TSP with time window and rejections (mTSPTWR), where\ncustomers who cannot be served before the deadline are subject to rejections.\nParticularly, in the proposed framework, a manager agent learns to divide\nmTSPTWR into sub-routing tasks by assigning customers to each vehicle via a\nGraph Isomorphism Network (GIN) based policy network. A worker agent learns to\nsolve sub-routing tasks by minimizing the cost in terms of both tour length and\nrejection rate for each vehicle, the maximum of which is then fed back to the\nmanager agent to learn better assignments. Experimental results demonstrate\nthat the proposed framework outperforms strong baselines in terms of higher\nsolution quality and shorter computation time. More importantly, the trained\nagents also achieve competitive performance for solving unseen larger\ninstances.\n", "date": "2022-09-13", "categories": []} {"id": "2209.05170", "title": "Resource Allocation to Agents with Restrictions: Maximizing Likelihood\n with Minimum Compromise", "abstract": " Many scenarios where agents with restrictions compete for resources can be\ncast as maximum matching problems on bipartite graphs. Our focus is on resource\nallocation problems where agents may have restrictions that make them\nincompatible with some resources. We assume that a Principle chooses a maximum\nmatching randomly so that each agent is matched to a resource with some\nprobability. Agents would like to improve their chances of being matched by\nmodifying their restrictions within certain limits. The Principle's goal is to\nadvise an unsatisfied agent to relax its restrictions so that the total cost of\nrelaxation is within a budget (chosen by the agent) and the increase in the\nprobability of being assigned a resource is maximized. We establish hardness\nresults for some variants of this budget-constrained maximization problem and\npresent algorithmic results for other variants. We experimentally evaluate our\nmethods on synthetic datasets as well as on two novel real-world datasets: a\nvacation activities dataset and a classrooms dataset.\n", "date": "2022-09-12", "categories": []} {"id": "2209.05187", "title": "Learning Obstacle-Avoiding Lattice Paths using Swarm Heuristics:\n Exploring the Bijection to Ordered Trees", "abstract": " Lattice paths are functional entities that model efficient navigation in\ndiscrete/grid maps. This paper presents a new scheme to generate collision-free\nlattice paths with utmost efficiency using the bijective property to rooted\nordered trees, rendering a one-dimensional search problem. Our computational\nstudies using ten state-of-the-art and relevant nature-inspired swarm\nheuristics in navigation scenarios with obstacles with convex and non-convex\ngeometry show the practical feasibility and efficiency in rendering\ncollision-free lattice paths. We believe our scheme may find use in devising\nfast algorithms for planning and combinatorial optimization in discrete maps.\n", "date": "2022-09-12", "categories": []} {"id": "2209.05206", "title": "A Differentiable Loss Function for Learning Heuristics in A*", "abstract": " Optimization of heuristic functions for the A* algorithm, realized by deep\nneural networks, is usually done by minimizing square root loss of estimate of\nthe cost to goal values. This paper argues that this does not necessarily lead\nto a faster search of A* algorithm since its execution relies on relative\nvalues instead of absolute ones. As a mitigation, we propose a L* loss, which\nupper-bounds the number of excessively expanded states inside the A* search.\nThe L* loss, when used in the optimization of state-of-the-art deep neural\nnetworks for automated planning in maze domains like Sokoban and maze with\nteleports, significantly improves the fraction of solved problems, the quality\nof founded plans, and reduces the number of expanded states to approximately\n50%\n", "date": "2022-09-12", "categories": []} {"id": "2208.13880", "title": "Inference and Optimization for Engineering and Physical Systems", "abstract": " The central object of this PhD thesis is known under different names in the\nfields of computer science and statistical mechanics. In computer science, it\nis called the Maximum Cut problem, one of the famous twenty-one Karp's original\nNP-hard problems, while the same object from Physics is called the Ising Spin\nGlass model. This model of a rich structure often appears as a reduction or\nreformulation of real-world problems from computer science, physics and\nengineering. However, solving this model exactly (finding the maximal cut or\nthe ground state) is likely to stay an intractable problem (unless $\\textit{P}\n= \\textit{NP}$) and requires the development of ad-hoc heuristics for every\nparticular family of instances.\n One of the bright and beautiful connections between discrete and continuous\noptimization is a Semidefinite Programming-based rounding scheme for Maximum\nCut. This procedure allows us to find a provably near-optimal solution;\nmoreover, this method is conjectured to be the best possible in polynomial\ntime. In the first two chapters of this thesis, we investigate local non-convex\nheuristics intended to improve the rounding scheme.\n In the last chapter of this thesis, we make one step further and aim to\ncontrol the solution of the problem we wanted to solve in previous chapters. We\nformulate a bi-level optimization problem over the Ising model where we want to\ntweak the interactions as little as possible so that the ground state of the\nresulting Ising model satisfies the desired criteria. This kind of problem\narises in pandemic modeling. We show that when the interactions are\nnon-negative, our bi-level optimization is solvable in polynomial time using\nconvex programming.\n", "date": "2022-08-29", "categories": []} {"id": "2208.12636", "title": "Four Algorithms for Correlation Clustering: A Survey", "abstract": " In the Correlation Clustering problem, we are given a set of objects with\npairwise similarity information. Our aim is to partition these objects into\nclusters that match this information as closely as possible. More specifically,\nthe pairwise information is given as a weighted graph $G$ with its edges\nlabelled as ``similar\" or ``dissimilar\" by a binary classifier. The goal is to\nproduce a clustering that minimizes the weight of ``disagreements\": the sum of\nthe weights of similar edges across clusters and dissimilar edges within\nclusters. In this exposition we focus on the case when $G$ is complete and\nunweighted. We explore four approximation algorithms for the Correlation\nClustering problem under this assumption. In particular, we describe the\nfollowing algorithms: (i) the $17429-$approximation algorithm by Bansal, Blum,\nand Chawla, (ii) the $4-$approximation algorithm by Charikar, Guruswami, and\nWirth (iii) the $3-$approximation algorithm by Ailon, Charikar, and Newman (iv)\nthe $2.06-$approximation algorithm by Chawla, Makarychev, Schramm, and\nYaroslavtsev.\n", "date": "2022-08-24", "categories": []} {"id": "2208.12191", "title": "Turning Mathematics Problems into Games: Reinforcement Learning and\n Gr\\\"obner bases together solve Integer Feasibility Problems", "abstract": " Can agents be trained to answer difficult mathematical questions by playing a\ngame? We consider the integer feasibility problem, a challenge of deciding\nwhether a system of linear equations and inequalities has a solution with\ninteger values. This is a famous NP-complete problem with applications in many\nareas of Mathematics and Computer Science. Our paper describes a novel\nalgebraic reinforcement learning framework that allows an agent to play a game\nequivalent to the integer feasibility problem. We explain how to transform the\ninteger feasibility problem into a game over a set of arrays with fixed margin\nsums. The game starts with an initial state (an array), and by applying a legal\nmove that leaves the margins unchanged, we aim to eventually reach a winning\nstate with zeros in specific positions. To win the game the player must find a\npath between the initial state and a final terminal winning state if one\nexists. Finding such a winning state is equivalent to solving the integer\nfeasibility problem. The key algebraic ingredient is a Gr\\\"obner basis of the\ntoric ideal for the underlying axial transportation polyhedron. The Gr\\\"obner\nbasis can be seen as a set of connecting moves (actions) of the game. We then\npropose a novel RL approach that trains an agent to predict moves in continuous\nspace to cope with the large size of action space. The continuous move is then\nprojected onto the set of legal moves so that the path always leads to valid\nstates. As a proof of concept we demonstrate in experiments that our agent can\nplay well the simplest version of our game for 2-way tables. Our work\nhighlights the potential to train agents to solve non-trivial mathematical\nqueries through contemporary machine learning methods used to train agents to\nplay games.\n", "date": "2022-08-25", "categories": []} {"id": "2208.10227", "title": "One Model, Any CSP: Graph Neural Networks as Fast Global Search\n Heuristics for Constraint Satisfaction", "abstract": " We propose a universal Graph Neural Network architecture which can be trained\nas an end-2-end search heuristic for any Constraint Satisfaction Problem (CSP).\nOur architecture can be trained unsupervised with policy gradient descent to\ngenerate problem specific heuristics for any CSP in a purely data driven\nmanner. The approach is based on a novel graph representation for CSPs that is\nboth generic and compact and enables us to process every possible CSP instance\nwith one GNN, regardless of constraint arity, relations or domain size. Unlike\nprevious RL-based methods, we operate on a global search action space and allow\nour GNN to modify any number of variables in every step of the stochastic\nsearch. This enables our method to properly leverage the inherent parallelism\nof GNNs. We perform a thorough empirical evaluation where we learn heuristics\nfor well known and important CSPs from random data, including graph coloring,\nMaxCut, 3-SAT and MAX-k-SAT. Our approach outperforms prior approaches for\nneural combinatorial optimization by a substantial margin. It can compete with,\nand even improve upon, conventional search heuristics on test instances that\nare several orders of magnitude larger and structurally more complex than those\nseen during training.\n", "date": "2022-08-22", "categories": []} {"id": "2208.06975", "title": "Rethinking Graph Neural Networks for the Graph Coloring Problem", "abstract": " Graph coloring, a classical and critical NP-hard problem, is the problem of\nassigning connected nodes as different colors as possible. However, we observe\nthat state-of-the-art GNNs are less successful in the graph coloring problem.\nWe analyze the reasons from two perspectives. First, most GNNs fail to\ngeneralize the task under homophily to heterophily, i.e., graphs where\nconnected nodes are assigned different colors. Second, GNNs are bounded by the\nnetwork depth, making them possible to be a local method, which has been\ndemonstrated to be non-optimal in Maximum Independent Set (MIS) problem. In\nthis paper, we focus on the aggregation-combine GNNs (AC-GNNs), a popular class\nof GNNs. We first define the power of AC-GNNs in the coloring problem as the\ncapability to assign nodes different colors. The definition is different with\nprevious one that is based on the assumption of homophily. We identify node\npairs that AC-GNNs fail to discriminate. Furthermore, we show that any AC-GNN\nis a local coloring method, and any local coloring method is non-optimal by\nexploring the limits of local methods over sparse random graphs, thereby\ndemonstrating the non-optimality of AC-GNNs due to its local property. We then\nprove the positive correlation between model depth and its coloring power.\nMoreover, we discuss the color equivariance of graphs to tackle some practical\nconstraints such as the pre-fixing constraints. Following the discussions\nabove, we summarize a series of rules a series of rules that make a GNN color\nequivariant and powerful in the coloring problem. Then, we propose a simple\nAC-GNN variation satisfying these rules. We empirically validate our\ntheoretical findings and demonstrate that our simple model substantially\noutperforms state-of-the-art heuristic algorithms in both quality and runtime.\n", "date": "2022-08-15", "categories": []} {"id": "2208.08620", "title": "Hybrid Learning with New Value Function for the Maximum Common Subgraph\n Problem", "abstract": " Maximum Common induced Subgraph (MCS) is an important NP-hard problem with\nwide real-world applications. Branch-and-Bound (BnB) is the basis of a class of\nefficient algorithms for MCS, consisting in successively selecting vertices to\nmatch and pruning when it is discovered that a solution better than the best\nsolution found so far does not exist. The method of selecting the vertices to\nmatch is essential for the performance of BnB. In this paper, we propose a new\nvalue function and a hybrid selection strategy used in reinforcement learning\nto define a new vertex selection method, and propose a new BnB algorithm,\ncalled McSplitDAL, for MCS. Extensive experiments show that McSplitDAL\nsignificantly improves the current best BnB algorithms, McSplit+LL and\nMcSplit+RL. An empirical analysis is also performed to illustrate why the new\nvalue function and the hybrid selection strategy are effective.\n", "date": "2022-08-18", "categories": []} {"id": "2208.08021", "title": "Streaming Adaptive Submodular Maximization", "abstract": " Many sequential decision making problems can be formulated as an adaptive\nsubmodular maximization problem. However, most of existing studies in this\nfield focus on pool-based setting, where one can pick items in any order, and\nthere have been few studies for the stream-based setting where items arrive in\nan arbitrary order and one must immediately decide whether to select an item or\nnot upon its arrival. In this paper, we introduce a new class of utility\nfunctions, semi-policywise submodular functions. We develop a series of\neffective algorithms to maximize a semi-policywise submodular function under\nthe stream-based setting.\n", "date": "2022-08-17", "categories": []} {"id": "2208.07123", "title": "Online 3D Bin Packing Reinforcement Learning Solution with Buffer", "abstract": " The 3D Bin Packing Problem (3D-BPP) is one of the most demanded yet\nchallenging problems in industry, where an agent must pack variable size items\ndelivered in sequence into a finite bin with the aim to maximize the space\nutilization. It represents a strongly NP-Hard optimization problem such that no\nsolution has been offered to date with high performance in space utilization.\nIn this paper, we present a new reinforcement learning (RL) framework for a\n3D-BPP solution for improving performance. First, a buffer is introduced to\nallow multi-item action selection. By increasing the degree of freedom in\naction selection, a more complex policy that results in better packing\nperformance can be derived. Second, we propose an agnostic data augmentation\nstrategy that exploits both bin item symmetries for improving sample\nefficiency. Third, we implement a model-based RL method adapted from the\npopular algorithm AlphaGo, which has shown superhuman performance in zero-sum\ngames. Our adaptation is capable of working in single-player and score based\nenvironments. In spite of the fact that AlphaGo versions are known to be\ncomputationally heavy, we manage to train the proposed framework with a single\nthread and GPU, while obtaining a solution that outperforms the\nstate-of-the-art results in space utilization.\n", "date": "2022-08-15", "categories": []} {"id": "2208.02867", "title": "Memetic algorithms for Spatial Partitioning problems", "abstract": " Spatial optimization problems (SOPs) are characterized by spatial\nrelationships governing the decision variables, objectives, and/or constraint\nfunctions. In this article, we focus on a specific type of SOP called spatial\npartitioning, which is a combinatorial problem due to the presence of discrete\nspatial units. Exact optimization methods do not scale with the size of the\nproblem, especially within practicable time limits. This motivated us to\ndevelop population-based metaheuristics for solving such SOPs. However, the\nsearch operators employed by these population-based methods are mostly designed\nfor real-parameter continuous optimization problems. For adapting these methods\nto SOPs, we apply domain knowledge in designing spatially-aware search\noperators for efficiently searching through the discrete search space while\npreserving the spatial constraints. To this end, we put forward a simple yet\neffective algorithm called swarm-based spatial memetic algorithm (SPATIAL) and\ntest it on the school (re)districting problem. Detailed experimental\ninvestigations are performed on real-world datasets to evaluate the performance\nof SPATIAL. Besides, ablation studies are performed to understand the role of\nthe individual components of SPATIAL. Additionally, we discuss how SPATIAL~is\nhelpful in the real-life planning process and its applicability to different\nscenarios and motivate future research directions.\n", "date": "2022-08-04", "categories": []} {"id": "2208.01222", "title": "Optimal and Bounded-Suboptimal Multi-Goal Task Assignment and Path\n Finding", "abstract": " We formalize and study the multi-goal task assignment and path finding\n(MG-TAPF) problem from theoretical and algorithmic perspectives. The MG-TAPF\nproblem is to compute an assignment of tasks to agents, where each task\nconsists of a sequence of goal locations, and collision-free paths for the\nagents that visit all goal locations of their assigned tasks in sequence.\nTheoretically, we prove that the MG-TAPF problem is NP-hard to solve optimally.\nWe present algorithms that build upon algorithmic techniques for the\nmulti-agent path finding problem and solve the MG-TAPF problem optimally and\nbounded-suboptimally. We experimentally compare these algorithms on a variety\nof different benchmark domains.\n", "date": "2022-08-02", "categories": []} {"id": "2208.00191", "title": "PUSH: a primal heuristic based on Feasibility PUmp and SHifting", "abstract": " This work describes PUSH, a primal heuristic combining Feasibility Pump and\nShifting. The main idea is to replace the rounding phase of the Feasibility\nPump with a suitable adaptation of the Shifting and other rounding heuristics.\nThe algorithm presents different strategies, depending on the nature of the\npartial rounding obtained. In particular, we distinguish when the partial\nsolution is feasible, infeasible with potential candidates, and infeasible\nwithout candidates. We used a threshold to indicate the percentage of variables\nto round with our algorithm and which other to round to the nearest integer.\nMost importantly, our algorithm tackles directly equality constraints without\nduplicating rows. We select the parameters of our algorithm on the 19 instances\nprovided for the Mip Competition 2022. Finally, we compared our approach to\nother start heuristics, like Simple Rounding, Rounding, Shifting, and\nFeasibility Pump on the first 800 MIPLIB2017 instances ordered by the number of\nnon-zeros.\n", "date": "2022-07-30", "categories": []} {"id": "2208.00202", "title": "Solving the vehicle routing problem with deep reinforcement learning", "abstract": " Recently, the applications of the methodologies of Reinforcement Learning\n(RL) to NP-Hard Combinatorial optimization problems have become a popular\ntopic. This is essentially due to the nature of the traditional combinatorial\nalgorithms, often based on a trial-and-error process. RL aims at automating\nthis process. At this regard, this paper focuses on the application of RL for\nthe Vehicle Routing Problem (VRP), a famous combinatorial problem that belongs\nto the class of NP-Hard problems. In this work, first, the problem is modeled\nas a Markov Decision Process (MDP) and then the PPO method (which belongs to\nthe Actor-Critic class of Reinforcement learning methods) is applied. In a\nsecond phase, the neural architecture behind the Actor and Critic has been\nestablished, choosing to adopt a neural architecture based on the Convolutional\nneural networks, both for the Actor and the Critic. This choice resulted in\neffectively addressing problems of different sizes. Experiments performed on a\nwide range of instances show that the algorithm has good generalization\ncapabilities and can reach good solutions in a short time. Comparisons between\nthe algorithm proposed and the state-of-the-art solver OR-TOOLS show that the\nlatter still outperforms the Reinforcement learning algorithm. However, there\nare future research perspectives, that aim to upgrade the current performance\nof the algorithm proposed.\n", "date": "2022-07-30", "categories": []} {"id": "2207.09860", "title": "Learning to Solve Soft-Constrained Vehicle Routing Problems with\n Lagrangian Relaxation", "abstract": " Vehicle Routing Problems (VRPs) in real-world applications often come with\nvarious constraints, therefore bring additional computational challenges to\nexact solution methods or heuristic search approaches. The recent idea to learn\nheuristic move patterns from sample data has become increasingly promising to\nreduce solution developing costs. However, using learning-based approaches to\naddress more types of constrained VRP remains a challenge. The difficulty lies\nin controlling for constraint violations while searching for optimal solutions.\nTo overcome this challenge, we propose a Reinforcement Learning based method to\nsolve soft-constrained VRPs by incorporating the Lagrangian relaxation\ntechnique and using constrained policy optimization. We apply the method on\nthree common types of VRPs, the Travelling Salesman Problem with Time Windows\n(TSPTW), the Capacitated VRP (CVRP) and the Capacitated VRP with Time Windows\n(CVRPTW), to show the generalizability of the proposed method. After comparing\nto existing RL-based methods and open-source heuristic solvers, we demonstrate\nits competitive performance in finding solutions with a good balance in travel\ndistance, constraint violations and inference speed.\n", "date": "2022-07-20", "categories": []} {"id": "2207.13701", "title": "Branch Ranking for Efficient Mixed-Integer Programming via Offline\n Ranking-based Policy Learning", "abstract": " Deriving a good variable selection strategy in branch-and-bound is essential\nfor the efficiency of modern mixed-integer programming (MIP) solvers. With MIP\nbranching data collected during the previous solution process, learning to\nbranch methods have recently become superior over heuristics. As\nbranch-and-bound is naturally a sequential decision making task, one should\nlearn to optimize the utility of the whole MIP solving process instead of being\nmyopic on each step. In this work, we formulate learning to branch as an\noffline reinforcement learning (RL) problem, and propose a long-sighted hybrid\nsearch scheme to construct the offline MIP dataset, which values the long-term\nutilities of branching decisions. During the policy training phase, we deploy a\nranking-based reward assignment scheme to distinguish the promising samples\nfrom the long-term or short-term view, and train the branching model named\nBranch Ranking via offline policy learning. Experiments on synthetic MIP\nbenchmarks and real-world tasks demonstrate that Branch Rankink is more\nefficient and robust, and can better generalize to large scales of MIP\ninstances compared to the widely used heuristics and state-of-the-art\nlearning-based branching models.\n", "date": "2022-07-26", "categories": []} {"id": "2207.13667", "title": "Unsupervised Training for Neural TSP Solver", "abstract": " There has been a growing number of machine learning methods for approximately\nsolving the travelling salesman problem. However, these methods often require\nsolved instances for training or use complex reinforcement learning approaches\nthat need a large amount of tuning. To avoid these problems, we introduce a\nnovel unsupervised learning approach. We use a relaxation of an integer linear\nprogram for TSP to construct a loss function that does not require correct\ninstance labels. With variable discretization, its minimum coincides with the\noptimal or near-optimal solution. Furthermore, this loss function is\ndifferentiable and thus can be used to train neural networks directly. We use\nour loss function with a Graph Neural Network and design controlled experiments\non both Euclidean and asymmetric TSP. Our approach has the advantage over\nsupervised learning of not requiring large labelled datasets. In addition, the\nperformance of our approach surpasses reinforcement learning for asymmetric TSP\nand is comparable to reinforcement learning for Euclidean instances. Our\napproach is also more stable and easier to train than reinforcement learning.\n", "date": "2022-07-27", "categories": []} {"id": "2207.12509", "title": "Optimizing Empty Container Repositioning and Fleet Deployment via\n Configurable Semi-POMDPs", "abstract": " With the continuous growth of the global economy and markets, resource\nimbalance has risen to be one of the central issues in real logistic scenarios.\nIn marine transportation, this trade imbalance leads to Empty Container\nRepositioning (ECR) problems. Once the freight has been delivered from an\nexporting country to an importing one, the laden will turn into empty\ncontainers that need to be repositioned to satisfy new goods requests in\nexporting countries. In such problems, the performance that any cooperative\nrepositioning policy can achieve strictly depends on the routes that vessels\nwill follow (i.e., fleet deployment). Historically, Operation Research (OR)\napproaches were proposed to jointly optimize the repositioning policy along\nwith the fleet of vessels. However, the stochasticity of future supply and\ndemand of containers, together with black-box and non-linear constraints that\nare present within the environment, make these approaches unsuitable for these\nscenarios. In this paper, we introduce a novel framework, Configurable\nSemi-POMDPs, to model this type of problems. Furthermore, we provide a\ntwo-stage learning algorithm, \"Configure & Conquer\" (CC), that first configures\nthe environment by finding an approximation of the optimal fleet deployment\nstrategy, and then \"conquers\" it by learning an ECR policy in this tuned\nenvironmental setting. We validate our approach in large and real-world\ninstances of the problem. Our experiments highlight that CC avoids the pitfalls\nof OR methods and that it is successful at optimizing both the ECR policy and\nthe fleet of vessels, leading to superior performance in world trade\nenvironments.\n", "date": "2022-07-25", "categories": []} {"id": "2207.11542", "title": "Annealed Training for Combinatorial Optimization on Graphs", "abstract": " The hardness of combinatorial optimization (CO) problems hinders collecting\nsolutions for supervised learning. However, learning neural networks for CO\nproblems is notoriously difficult in lack of the labeled data as the training\nis easily trapped at local optima. In this work, we propose a simple but\neffective annealed training framework for CO problems. In particular, we\ntransform CO problems into unbiased energy-based models (EBMs). We carefully\nselected the penalties terms so as to make the EBMs as smooth as possible. Then\nwe train graph neural networks to approximate the EBMs. To prevent the training\nfrom being stuck at local optima near the initialization, we introduce an\nannealed loss function. An experimental evaluation demonstrates that our\nannealed training framework obtains substantial improvements. In four types of\nCO problems, our method achieves performance substantially better than other\nunsupervised neural methods on both synthetic and real-world graphs.\n", "date": "2022-07-23", "categories": []} {"id": "2207.11122", "title": "Solving the Batch Stochastic Bin Packing Problem in Cloud: A\n Chance-constrained Optimization Approach", "abstract": " This paper investigates a critical resource allocation problem in the first\nparty cloud: scheduling containers to machines. There are tens of services and\neach service runs a set of homogeneous containers with dynamic resource usage;\ncontainers of a service are scheduled daily in a batch fashion. This problem\ncan be naturally formulated as Stochastic Bin Packing Problem (SBPP). However,\ntraditional SBPP research often focuses on cases of empty machines, whose\nobjective, i.e., to minimize the number of used machines, is not well-defined\nfor the more common reality with nonempty machines. This paper aims to close\nthis gap. First, we define a new objective metric, Used Capacity at Confidence\n(UCaC), which measures the maximum used resources at a probability and is\nproved to be consistent for both empty and nonempty machines, and reformulate\nthe SBPP under chance constraints. Second, by modeling the container resource\nusage distribution in a generative approach, we reveal that UCaC can be\napproximated with Gaussian, which is verified by trace data of real-world\napplications. Third, we propose an exact solver by solving the equivalent\ncutting stock variant as well as two heuristics-based solvers -- UCaC best fit,\nbi-level heuristics. We experimentally evaluate these solvers on both synthetic\ndatasets and real application traces, demonstrating our methodology's advantage\nover traditional SBPP optimal solver minimizing the number of used machines,\nwith a low rate of resource violations.\n", "date": "2022-07-20", "categories": []} {"id": "2207.09588", "title": "New Auction Algorithms for Path Planning, Network Transport, and\n Reinforcement Learning", "abstract": " We consider some classical optimization problems in path planning and network\ntransport, and we introduce new auction-based algorithms for their optimal and\nsuboptimal solution. The algorithms are based on mathematical ideas that are\nrelated to competitive bidding by persons for objects and the attendant market\nequilibrium, which underlie auction processes. However, the starting point of\nour algorithms is different, namely weighted and unweighted path construction\nin directed graphs, rather than assignment of persons to objects. The new\nalgorithms have several potential advantages over existing methods: they are\nempirically faster in some important contexts, such as max-flow, they are\nwell-suited for on-line replanning, and they can be adapted to distributed\nasynchronous operation. Moreover, they allow arbitrary initial prices, without\ncomplementary slackness restrictions, and thus are better-suited to take\nadvantage of reinforcement learning methods that use off-line training with\ndata, as well as on-line training during real-time operation. The new\nalgorithms may also find use in reinforcement learning contexts involving\napproximation, such as multistep lookahead and tree search schemes, and/or\nrollout algorithms.\n", "date": "2022-07-19", "categories": []} {"id": "2207.07690", "title": "Support Vector Machines with the Hard-Margin Loss: Optimal Training via\n Combinatorial Benders' Cuts", "abstract": " The classical hinge-loss support vector machines (SVMs) model is sensitive to\noutlier observations due to the unboundedness of its loss function. To\ncircumvent this issue, recent studies have focused on non-convex loss\nfunctions, such as the hard-margin loss, which associates a constant penalty to\nany misclassified or within-margin sample. Applying this loss function yields\nmuch-needed robustness for critical applications but it also leads to an\nNP-hard model that makes training difficult, since current exact optimization\nalgorithms show limited scalability, whereas heuristics are not able to find\nhigh-quality solutions consistently. Against this background, we propose new\ninteger programming strategies that significantly improve our ability to train\nthe hard-margin SVM model to global optimality. We introduce an iterative\nsampling and decomposition approach, in which smaller subproblems are used to\nseparate combinatorial Benders' cuts. Those cuts, used within a branch-and-cut\nalgorithm, permit to converge much more quickly towards a global optimum.\nThrough extensive numerical analyses on classical benchmark data sets, our\nsolution algorithm solves, for the first time, 117 new data sets to optimality\nand achieves a reduction of 50% in the average optimality gap for the hardest\ndatasets of the benchmark.\n", "date": "2022-07-15", "categories": []} {"id": "2207.08398", "title": "Bayesian Optimization for Macro Placement", "abstract": " Macro placement is the problem of placing memory blocks on a chip canvas. It\ncan be formulated as a combinatorial optimization problem over sequence pairs,\na representation which describes the relative positions of macros. Solving this\nproblem is particularly challenging since the objective function is expensive\nto evaluate. In this paper, we develop a novel approach to macro placement\nusing Bayesian optimization (BO) over sequence pairs. BO is a machine learning\ntechnique that uses a probabilistic surrogate model and an acquisition function\nthat balances exploration and exploitation to efficiently optimize a black-box\nobjective function. BO is more sample-efficient than reinforcement learning and\ntherefore can be used with more realistic objectives. Additionally, the ability\nto learn from data and adapt the algorithm to the objective function makes BO\nan appealing alternative to other black-box optimization methods such as\nsimulated annealing, which relies on problem-dependent heuristics and\nparameter-tuning. We benchmark our algorithm on the fixed-outline macro\nplacement problem with the half-perimeter wire length objective and demonstrate\ncompetitive performance.\n", "date": "2022-07-18", "categories": []} {"id": "2207.03876", "title": "Reinforced Lin-Kernighan-Helsgaun Algorithms for the Traveling Salesman\n Problems", "abstract": " TSP is a classical NP-hard combinatorial optimization problem with many\npractical variants. LKH is one of the state-of-the-art local search algorithms\nfor the TSP. LKH-3 is a powerful extension of LKH that can solve many TSP\nvariants. Both LKH and LKH-3 associate a candidate set to each city to improve\nthe efficiency, and have two different methods, $\\alpha$-measure and POPMUSIC,\nto decide the candidate sets. In this work, we first propose a Variable\nStrategy Reinforced LKH (VSR-LKH) algorithm, which incorporates three\nreinforcement learning methods (Q-learning, Sarsa, Monte Carlo) with LKH, for\nthe TSP. We further propose a new algorithm called VSR-LKH-3 that combines the\nvariable strategy reinforcement learning method with LKH-3 for typical TSP\nvariants, including the TSP with time windows (TSPTW) and Colored TSP (CTSP).\nThe proposed algorithms replace the inflexible traversal operations in LKH and\nLKH-3 and let the algorithms learn to make a choice at each search step by\nreinforcement learning. Both LKH and LKH-3, with either $\\alpha$-measure or\nPOPMUSIC, can be significantly improved by our methods. Extensive experiments\non 236 widely-used TSP benchmarks with up to 85,900 cities demonstrate the\nexcellent performance of VSR-LKH. VSR-LKH-3 also significantly outperforms the\nstate-of-the-art heuristics for TSPTW and CTSP.\n", "date": "2022-07-08", "categories": []} {"id": "2207.07212", "title": "Attention, Filling in The Gaps for Generalization in Routing Problems", "abstract": " Machine Learning (ML) methods have become a useful tool for tackling vehicle\nrouting problems, either in combination with popular heuristics or as\nstandalone models. However, current methods suffer from poor generalization\nwhen tackling problems of different sizes or different distributions. As a\nresult, ML in vehicle routing has witnessed an expansion phase with new\nmethodologies being created for particular problem instances that become\ninfeasible at larger problem sizes.\n This paper aims at encouraging the consolidation of the field through\nunderstanding and improving current existing models, namely the attention model\nby Kool et al. We identify two discrepancy categories for VRP generalization.\nThe first is based on the differences that are inherent to the problems\nthemselves, and the second relates to architectural weaknesses that limit the\nmodel's ability to generalize. Our contribution becomes threefold: We first\ntarget model discrepancies by adapting the Kool et al. method and its loss\nfunction for Sparse Dynamic Attention based on the alpha-entmax activation. We\nthen target inherent differences through the use of a mixed instance training\nmethod that has been shown to outperform single instance training in certain\nscenarios. Finally, we introduce a framework for inference level data\naugmentation that improves performance by leveraging the model's lack of\ninvariance to rotation and dilation changes.\n", "date": "2022-07-14", "categories": []} {"id": "2207.01411", "title": "The Neural-Prediction based Acceleration Algorithm of Column Generation\n for Graph-Based Set Covering Problems", "abstract": " Set covering problem is an important class of combinatorial optimization\nproblems, which has been widely applied and studied in many fields. In this\npaper, we propose an improved column generation algorithm with neural\nprediction (CG-P) for solving graph-based set covering problems. We leverage a\ngraph neural network based neural prediction model to predict the probability\nto be included in the final solution for each edge. Our CG-P algorithm\nconstructs a reduced graph that only contains the edges with higher predicted\nprobability, and this graph reduction process significantly speeds up the\nsolution process. We evaluate the CG-P algorithm on railway crew scheduling\nproblems and it outperforms the baseline column generation algorithm. We\nprovide two solution modes for our CG-P algorithm. In the optimal mode, we can\nobtain a solution with an optimality guarantee while reducing the time cost to\n63.12%. In the fast mode, we can obtain a sub-optimal solution with a 7.62%\noptimality gap in only 2.91% computation time.\n", "date": "2022-07-04", "categories": []} {"id": "2207.04520", "title": "Local Area Routes for Vehicle Routing Problems", "abstract": " We consider an approach for improving the efficiency of column generation\n(CG) methods for solving vehicle routing problems. We introduce Local Area (LA)\nroute relaxations, an alternative/complement to the commonly used ng-route\nrelaxations and Decremental State Space Relaxations (DSSR) inside of CG\nformulations. LA routes are a subset of ng-routes and a super-set of elementary\nroutes. Normally, the pricing stage of CG must produce elementary routes, which\nare routes without repeated customers, using processes which can be\ncomputationally expensive. Non-elementary routes visit at least one customer\nmore than once, creating a cycle. LA routes relax the constraint of being an\nelementary route in such a manner as to permit efficient pricing. LA routes are\nbest understood in terms of ng-route relaxations. Ng-routes are routes which\nare permitted to have non-localized cycles in space; this means that at least\none intermediate customer (called a breaker) in the cycle must consider the\nstarting customer in the cycle to be spatially far away. LA routes are\ndescribed using a set of special indexes corresponding to customers on the\nroute ordered from the start to the end of the route. LA route relaxations\nfurther restrict the set of permitted cycles beyond that of ng-routes by\nadditionally enforcing that the breaker must be a located at a special index\nwhere the set of special indexes is defined recursively as follows. The first\nspecial index in the route is at index 1 meaning that it is associated with the\nfirst customer in the route. The k'th special index corresponds to the first\ncustomer after the k-1'th special index, that is not considered to be a\nneighbor of (considered spatially far from) the customer located at the k-1'th\nspecial index. We demonstrate that LA route relaxations can significantly\nimprove the computational speed of pricing when compared to the standard DSSR.\n", "date": "2022-07-10", "categories": []} {"id": "2207.02937", "title": "Learning Optimal Solutions via an LSTM-Optimization Framework", "abstract": " In this study, we present a deep learning-optimization framework to tackle\ndynamic mixed-integer programs. Specifically, we develop a bidirectional Long\nShort Term Memory (LSTM) framework that can process information forward and\nbackward in time to learn optimal solutions to sequential decision-making\nproblems. We demonstrate our approach in predicting the optimal decisions for\nthe single-item capacitated lot-sizing problem (CLSP), where a binary variable\ndenotes whether to produce in a period or not. Due to the dynamic nature of the\nproblem, the CLSP can be treated as a sequence labeling task where a recurrent\nneural network can capture the problem's temporal dynamics. Computational\nresults show that our LSTM-Optimization (LSTM-Opt) framework significantly\nreduces the solution time of benchmark CLSP problems without much loss in\nfeasibility and optimality. For example, the predictions at the 85\\% level\nreduce the CPLEX solution time by a factor of 9 on average for over 240,000\ntest instances with an optimality gap of less than 0.05\\% and 0.4\\%\ninfeasibility in the test set. Also, models trained using shorter planning\nhorizons can successfully predict the optimal solution of the instances with\nlonger planning horizons. For the hardest data set, the LSTM predictions at the\n25\\% level reduce the solution time of 70 CPU hours to less than 2 CPU minutes\nwith an optimality gap of 0.8\\% and without any infeasibility. The LSTM-Opt\nframework outperforms classical ML algorithms, such as the logistic regression\nand random forest, in terms of the solution quality, and exact approaches, such\nas the ($\\ell$, S) and dynamic programming-based inequalities, with respect to\nthe solution time improvement. Our machine learning approach could be\nbeneficial in tackling sequential decision-making problems similar to CLSP,\nwhich need to be solved repetitively, frequently, and in a fast manner.\n", "date": "2022-07-06", "categories": []} {"id": "2207.01752", "title": "Plan Execution for Multi-Agent Path Finding with Indoor Quadcopters", "abstract": " We study the planning and acting phase for the problem of multi-agent path\nfinding (MAPF) in this paper. MAPF is a problem of navigating agents from their\nstart positions to specified individual goal positions so that agents do not\ncollide with each other. Specifically we focus on executing MAPF plans with a\ngroup of Crazyflies, small indoor quadcopters . We show how to modify the\nexisting continuous time conflict-based search algorithm (CCBS) to produce\nplans that are suitable for execution with the quadcopters. The acting phase\nuses the the Loco positioning system to check if the plan is executed\ncorrectly. Our finding is that the CCBS algorithm allows for extensions that\ncan produce safe plans for quadcopters, namely cylindrical protection zone\naround each quadcopter can be introduced at the planning level.\n", "date": "2022-07-05", "categories": []} {"id": "2207.02641", "title": "Reforming an Envy-Free Matching", "abstract": " We consider the problem of reforming an envy-free matching when each agent is\nassigned a single item. Given an envy-free matching, we consider an operation\nto exchange the item of an agent with an unassigned item preferred by the agent\nthat results in another envy-free matching. We repeat this operation as long as\nwe can. We prove that the resulting envy-free matching is uniquely determined\nup to the choice of an initial envy-free matching, and can be found in\npolynomial time. We call the resulting matching a reformist envy-free matching,\nand then we study a shortest sequence to obtain the reformist envy-free\nmatching from an initial envy-free matching. We prove that a shortest sequence\nis computationally hard to obtain even when each agent accepts at most four\nitems and each item is accepted by at most three agents. On the other hand, we\ngive polynomial-time algorithms when each agent accepts at most three items or\neach item is accepted by at most two agents. Inapproximability and\nfixed-parameter (in)tractability are also discussed.\n", "date": "2022-07-06", "categories": []} {"id": "2207.02659", "title": "Planning Courses for Student Success at the American College of Greece", "abstract": " We model the problem of optimizing the schedule of courses a student at the\nAmerican College of Greece will need to take to complete their studies. We\nmodel all constraints set forth by the institution and the department, so that\nwe guarantee the validity of all produced schedules. We formulate several\ndifferent objectives to optimize in the resulting schedule, including fastest\ncompletion time, course difficulty balance, and so on, with a very important\nobjective our model is capable of capturing being the maximization of the\nexpected student GPA given their performance on passed courses using Machine\nLearning and Data Mining techniques. All resulting problems are Mixed Integer\nLinear Programming problems with a number of binary variables that is in the\norder of the maximum number of terms times the number of courses available for\nthe student to take. The resulting Mathematical Programming problem is always\nsolvable by the GUROBI solver in less than 10 seconds on a modern commercial\noff-the-self PC, whereas the manual process that was installed before used to\ntake department heads that are designated as student advisors more than one\nhour of their time for every student and was resulting in sub-optimal schedules\nas measured by the objectives set forth.\n", "date": "2022-06-16", "categories": []} {"id": "2207.01693", "title": "Intelligent Exploration of Solution Spaces Exemplified by Industrial\n Reconfiguration Management", "abstract": " Many decision-making approaches rely on the exploration of solution spaces\nwith regards to specified criteria. However, in complex environments,\nbrute-force exploration strategies are usually not feasible. As an alternative,\nwe propose the combination of an exploration task's vertical sub-division into\nlayers representing different sequentially interdependent sub-problems of the\nparamount problem and a horizontal sub-division into self-sustained solution\nsub-spaces. In this paper, we present a universal methodology for the\nintelligent exploration of solution spaces and derive a use-case specific\nexample from the field of reconfiguration management in industry 4.0.\n", "date": "2022-07-04", "categories": []} {"id": "2207.02087", "title": "Learning to Accelerate Approximate Methods for Solving Integer\n Programming via Early Fixing", "abstract": " Integer programming (IP) is an important and challenging problem. Approximate\nmethods have shown promising performance on both effectiveness and efficiency\nfor solving the IP problem. However, we observed that a large fraction of\nvariables solved by some iterative approximate methods fluctuate around their\nfinal converged discrete states in very long iterations. Inspired by this\nobservation, we aim to accelerate these approximate methods by early fixing\nthese fluctuated variables to their converged states while not significantly\nharming the solution accuracy. To this end, we propose an early fixing\nframework along with the approximate method. We formulate the whole early\nfixing process as a Markov decision process, and train it using imitation\nlearning. A policy network will evaluate the posterior probability of each free\nvariable concerning its discrete candidate states in each block of iterations.\nSpecifically, we adopt the powerful multi-headed attention mechanism in the\npolicy network. Extensive experiments on our proposed early fixing framework\nare conducted to three different IP applications: constrained linear\nprogramming, MRF energy minimization and sparse adversarial attack. The former\none is linear IP problem, while the latter two are quadratic IP problems. We\nextend the problem scale from regular size to significantly large size. The\nextensive experiments reveal the competitiveness of our early fixing framework:\nthe runtime speeds up significantly, while the solution quality does not\ndegrade much, even in some cases it is available to obtain better solutions.\nOur proposed early fixing framework can be regarded as an acceleration\nextension of ADMM methods for solving integer programming. The source codes are\navailable at \\url{https://github.com/SCLBD/Accelerated-Lpbox-ADMM}.\n", "date": "2022-07-05", "categories": []} {"id": "2207.00745", "title": "Scheduling Planting Time Through Developing an Optimization Model and\n Analysis of Time Series Growing Degree Units", "abstract": " Producing higher-quality crops within shortened breeding cycles ensures\nglobal food availability and security, but this improvement intensifies\nlogistical and productivity challenges for seed industries in the year-round\nbreeding process due to the storage limitations. In the 2021 Syngenta crop\nchallenge in analytics, Syngenta raised the problem to design an optimization\nmodel for the planting time scheduling in the 2020 year-round breeding process\nso that there is a consistent harvest quantity each week. They released a\ndataset that contained 2569 seed populations with their planting windows,\nrequired growing degree units for harvesting, and their harvest quantities at\ntwo sites. To address this challenge, we developed a new framework that\nconsists of a weather time series model and an optimization model to schedule\nthe planting time. A deep recurrent neural network was designed to predict the\nweather into the future, and a Gaussian process model on top of the time-series\nmodel was developed to model the uncertainty of forecasted weather. The\nproposed optimization models also scheduled the seed population's planting time\nat the fewest number of weeks with a more consistent weekly harvest quantity.\nUsing the proposed optimization models can decrease the required capacity by\n69% at site 0 and up to 51% at site 1 compared to the original planting time.\n", "date": "2022-07-02", "categories": []} {"id": "2207.00919", "title": "An AlphaZero-Inspired Approach to Solving Search Problems", "abstract": " AlphaZero and its extension MuZero are computer programs that use\nmachine-learning techniques to play at a superhuman level in chess, go, and a\nfew other games. They achieved this level of play solely with reinforcement\nlearning from self-play, without any domain knowledge except the game rules. It\nis a natural idea to adapt the methods and techniques used in AlphaZero for\nsolving search problems such as the Boolean satisfiability problem (in its\nsearch version). Given a search problem, how to represent it for an\nAlphaZero-inspired solver? What are the \"rules of solving\" for this search\nproblem? We describe possible representations in terms of easy-instance solvers\nand self-reductions, and we give examples of such representations for the\nsatisfiability problem. We also describe a version of Monte Carlo tree search\nadapted for search problems.\n", "date": "2022-07-02", "categories": []} {"id": "2207.01239", "title": "Satellite downlink scheduling under breakpoint resume mode", "abstract": " A novel problem called satellite downlink scheduling problem (SDSP) under\nbreakpoint resume mode (SDSP-BRM) is studied in our paper. Compared to the\ntraditional SDSP where an imaging data has to be completely downloaded at one\ntime, SDSP-BRM allows the data of an imaging data be broken into a number of\npieces which can be downloaded in different playback windows. By analyzing the\ncharacteristics of SDSP-BRM, we first propose a mixed integer programming model\nfor its formulation and then prove the NP-hardness of SDSP-BRM. To solve the\nproblem, we design a simple and effective heuristic algorithm (SEHA) where a\nnumber of problem-tailored move operators are proposed for local searching.\nNumerical results on a set of well-designed scenarios demonstrate the\nefficiency of the proposed algorithm in comparison to the general purpose CPLEX\nsolver. We conduct additional experiments to shed light on the impact of the\nsegmental strategy on the overall performance of the proposed SEHA.\n", "date": "2022-07-04", "categories": []} {"id": "2207.01250", "title": "Three multi-objective memtic algorithms for observation scheduling\n problem of active-imaging AEOS", "abstract": " Observation scheduling problem for agile earth observation satellites\n(OSPFAS) plays a critical role in management of agile earth observation\nsatellites (AEOSs). Active imaging enriches the extension of OSPFAS, we call\nthe novel problem as observation scheduling problem for AEOS with variable\nimage duration (OSWVID). A cumulative image quality and a detailed energy\nconsumption is proposed to build OSWVID as a bi-objective optimization model.\nThree multi-objective memetic algorithms, PD+NSGA-II, LA+NSGA-II and\nALNS+NSGA-II, are then designed to solve OSWVID. Considering the heuristic\nknowledge summarized in our previous research, several operators are designed\nfor improving these three algorithms respectively. Based on existing instances,\nwe analyze the critical parameters optimization, operators evolution, and\nefficiency of these three algorithms according to extensive simulation\nexperiments.\n", "date": "2022-07-04", "categories": []} {"id": "2207.01257", "title": "Multi-strip observation scheduling problem for ac-tive-imaging agile\n earth observation satellites", "abstract": " Active-imaging agile earth observation satellite (AI-AEOS) is a new\ngeneration agile earth observation satellite (AEOS). With renewed capabilities\nin observation and active im-aging, AI-AEOS improves upon the observation\ncapabilities of AEOS and provide additional ways to observe ground targets.\nThis however makes the observation scheduling problem for these agile earth\nobservation satellite more complex, especially when considering multi-strip\nground targets. In this paper, we investigate the multi-strip observation\nscheduling problem for an active-image agile earth observation satellite\n(MOSP). A bi-objective optimization model is presented for MOSP along with an\nadaptive bi-objective memetic algorithm which integrates the combined power of\nan adaptive large neighborhood search algorithm (ALNS) and a nondominated\nsorting genetic algorithm II (NSGA-II). Results of extensive computa-tional\nexperiments are presented which disclose that ALNS and NSGA-II when worked in\nunison produced superior outcomes. Our model is more versatile than existing\nmodels and provide enhanced capabilities in applied problem solving.\n", "date": "2022-07-04", "categories": []} {"id": "2207.00168", "title": "An adaptive bi-objective optimization algorithm for the satellite image\n data downlink scheduling problem considering request split", "abstract": " The satellite image data downlink scheduling problem (SIDSP) is well studied\nin literature for traditional satellites. With recent developments in satellite\ntechnology, SIDSP for modern satellites became more complicated, adding new\ndimensions of complexities and additional opportunities for the effective use\nof the satellite. In this paper, we introduce the dynamic two-phase satellite\nimage data downlink scheduling problem (D-SIDSP) which combines two interlinked\noperations of image data segmentation and image data downlink, in a dynamic\nway, and thereby offering additional modelling flexibility and renewed\ncapabilities. D-SIDSP is formulated as a bi-objective problem of optimizing the\nimage data transmission rate and the service-balance degree. Harnessing the\npower of an adaptive large neighborhood search algorithm (ALNS) with a\nnondominated sorting genetic algorithm II (NSGA-II), an adaptive bi-objective\nmemetic algorithm, ALNS+NSGA-II, is developed to solve D-SIDSP. Results of\nextensive computational experiments carried out using benchmark instances are\nalso presented. Our experimental results disclose that the algorithm\nALNS+NSGA-II is a viable alternative to solve D-SIDSP more efficiently and\ndemonstrates superior outcomes based on various performance metrics. The paper\nalso offers new benchmark instances for D-SIDSP that can be used in future\nresearch works on the topic.\n", "date": "2022-06-28", "categories": []} {"id": "2206.12413", "title": "A Novel Multi-Agent Scheduling Mechanism for Adaptation of Production\n Plans in Case of Supply Chain Disruptions", "abstract": " Manufacturing companies typically use sophisticated production planning\nsystems optimizing production steps, often delivering near-optimal solutions.\nAs a downside for delivering a near-optimal schedule, planning systems have\nhigh computational demands resulting in hours of computation. Under normal\ncircumstances this is not issue if there is enough buffer time before\nimplementation of the schedule (e.g. at night for the next day). However, in\ncase of unexpected disruptions such as delayed part deliveries or defectively\nmanufactured goods, the planned schedule may become invalid and swift\nreplanning becomes necessary. Such immediate replanning is unsuited for\nexisting optimal planners due to the computational requirements. This paper\nproposes a novel solution that can effectively and efficiently perform\nreplanning in case of different types of disruptions using an existing plan.\nThe approach is based on the idea to adhere to the existing schedule as much as\npossible, adapting it based on limited local changes. For that purpose an\nagent-based scheduling mechanism has been devised, in which agents represent\nmaterials and production sites and use local optimization techniques and\nnegotiations to generate an adapted (sufficient, but non-optimal) schedule. The\napproach has been evaluated using real production data from Huawei, showing\nthat efficient schedules are produced in short time. The system has been\nimplemented as proof of concept and is currently reimplemented and transferred\nto a production system based on the Jadex agent platform.\n", "date": "2022-06-23", "categories": []} {"id": "2206.13414", "title": "Learning To Cut By Looking Ahead: Cutting Plane Selection via Imitation\n Learning", "abstract": " Cutting planes are essential for solving mixed-integer linear problems\n(MILPs), because they facilitate bound improvements on the optimal solution\nvalue. For selecting cuts, modern solvers rely on manually designed heuristics\nthat are tuned to gauge the potential effectiveness of cuts. We show that a\ngreedy selection rule explicitly looking ahead to select cuts that yield the\nbest bound improvement delivers strong decisions for cut selection - but is too\nexpensive to be deployed in practice. In response, we propose a new neural\narchitecture (NeuralCut) for imitation learning on the lookahead expert. Our\nmodel outperforms standard baselines for cut selection on several synthetic\nMILP benchmarks. Experiments with a B&C solver for neural network verification\nfurther validate our approach, and exhibit the potential of learning methods in\nthis setting.\n", "date": "2022-06-27", "categories": []} {"id": "2206.11040", "title": "Penalty Weights in QUBO Formulations: Permutation Problems", "abstract": " Optimisation algorithms designed to work on quantum computers or other\nspecialised hardware have been of research interest in recent years. Many of\nthese solver can only optimise problems that are in binary and quadratic form.\nQuadratic Unconstrained Binary Optimisation (QUBO) is therefore a common\nformulation used by these solvers.\n There are many combinatorial optimisation problems that are naturally\nrepresented as permutations e.g., travelling salesman problem. Encoding\npermutation problems using binary variables however presents some challenges.\nMany QUBO solvers are single flip solvers, it is therefore possible to generate\nsolutions that cannot be decoded to a valid permutation. To create bias towards\ngenerating feasible solutions, we use penalty weights. The process of setting\nstatic penalty weights for various types of problems is not trivial. This is\nbecause values that are too small will lead to infeasible solutions being\nreturned by the solver while values that are too large may lead to slower\nconvergence. In this study, we explore some methods of setting penalty weights\nwithin the context of QUBO formulations. We propose new static methods of\ncalculating penalty weights which lead to more promising results than existing\nmethods.\n", "date": "2022-06-20", "categories": []} {"id": "2206.11210", "title": "Constant-Factor Approximation Algorithms for Socially Fair\n $k$-Clustering", "abstract": " We study approximation algorithms for the socially fair $(\\ell_p,\nk)$-clustering problem with $m$ groups, whose special cases include the\nsocially fair $k$-median ($p=1$) and socially fair $k$-means ($p=2$) problems.\nWe present (1) a polynomial-time $(5+2\\sqrt{6})^p$-approximation with at most\n$k+m$ centers (2) a $(5+2\\sqrt{6}+\\epsilon)^p$-approximation with $k$ centers\nin time $n^{2^{O(p)}\\cdot m^2}$, and (3) a $(15+6\\sqrt{6})^p$ approximation\nwith $k$ centers in time $k^{m}\\cdot\\text{poly}(n)$. The first result is\nobtained via a refinement of the iterative rounding method using a sequence of\nlinear programs. The latter two results are obtained by converting a solution\nwith up to $k+m$ centers to one with $k$ centers using sparsification methods\nfor (2) and via an exhaustive search for (3). We also compare the performance\nof our algorithms with existing bicriteria algorithms as well as exactly $k$\ncenter approximation algorithms on benchmark datasets, and find that our\nalgorithms also outperform existing methods in practice.\n", "date": "2022-06-22", "categories": []} {"id": "2206.01857", "title": "Design and Implementation of an Heuristic-Enhanced Branch-and-Bound\n Solver for MILP", "abstract": " We present a solver for Mixed Integer Programs (MIP) developed for the MIP\ncompetition 2022. Given the 10 minutes bound on the computational time\nestablished by the rules of the competition, our method focuses on finding a\nfeasible solution and improves it through a Branch-and-Bound algorithm. Another\nrule of the competition allows the use of up to 8 threads. Each thread is given\na different primal heuristic, which has been tuned by hyper-parameters, to find\na feasible solution. In every thread, once a feasible solution is found, we\nstop and we use a Branch-and-Bound method, embedded with local search\nheuristics, to ameliorate the incumbent solution. The three variants of the\nDiving heuristic that we implemented manage to find a feasible solution for 10\ninstances of the training data set. These heuristics are the best performing\namong the heuristics that we implemented. Our Branch-and-Bound algorithm is\neffective on a small portion of the training data set, and it manages to find\nan incumbent feasible solution for an instance that we could not solve with the\nDiving heuristics. Overall, our combined methods, when implemented with\nextensive computational power, can solve 11 of the 19 problems of the training\ndata set within the time limit. Our submission to the MIP competition was\nawarded the \"Outstanding Student Submission\" honorable mention.\n", "date": "2022-06-04", "categories": []} {"id": "2206.10178", "title": "Large region targets observation scheduling by multiple satellites using\n resampling particle swarm optimization", "abstract": " The last decades have witnessed a rapid increase of Earth observation\nsatellites (EOSs), leading to the increasing complexity of EOSs scheduling. On\naccount of the widespread applications of large region observation, this paper\naims to address the EOSs observation scheduling problem for large region\ntargets. A rapid coverage calculation method employing a projection reference\nplane and a polygon clipping technique is first developed. We then formulate a\nnonlinear integer programming model for the scheduling problem, where the\nobjective function is calculated based on the developed coverage calculation\nmethod. A greedy initialization-based resampling particle swarm optimization\n(GI-RPSO) algorithm is proposed to solve the model. The adopted greedy\ninitialization strategy and particle resampling method contribute to generating\nefficient and effective solutions during the evolution process. In the end,\nextensive experiments are conducted to illustrate the effectiveness and\nreliability of the proposed method. Compared to the traditional particle swarm\noptimization and the widely used greedy algorithm, the proposed GI-RPSO can\nimprove the scheduling result by 5.42% and 15.86%, respectively.\n", "date": "2022-06-21", "categories": []} {"id": "2206.06618", "title": "Solving the capacitated vehicle routing problem with timing windows\n using rollouts and MAX-SAT", "abstract": " The vehicle routing problem is a well known class of NP-hard combinatorial\noptimisation problems in literature. Traditional solution methods involve\neither carefully designed heuristics, or time-consuming metaheuristics. Recent\nwork in reinforcement learning has been a promising alternative approach, but\nhas found it difficult to compete with traditional methods in terms of solution\nquality. This paper proposes a hybrid approach that combines reinforcement\nlearning, policy rollouts, and a satisfiability solver to enable a tunable\ntradeoff between computation times and solution quality. Results on a popular\npublic data set show that the algorithm is able to produce solutions closer to\noptimal levels than existing learning based approaches, and with shorter\ncomputation times than meta-heuristics. The approach requires minimal design\neffort and is able to solve unseen problems of arbitrary scale without\nadditional training. Furthermore, the methodology is generalisable to other\ncombinatorial optimisation problems.\n", "date": "2022-06-14", "categories": []} {"id": "2206.06965", "title": "Deep Reinforcement Learning for Exact Combinatorial Optimization:\n Learning to Branch", "abstract": " Branch-and-bound is a systematic enumerative method for combinatorial\noptimization, where the performance highly relies on the variable selection\nstrategy. State-of-the-art handcrafted heuristic strategies suffer from\nrelatively slow inference time for each selection, while the current machine\nlearning methods require a significant amount of labeled data. We propose a new\napproach for solving the data labeling and inference latency issues in\ncombinatorial optimization based on the use of the reinforcement learning (RL)\nparadigm. We use imitation learning to bootstrap an RL agent and then use\nProximal Policy Optimization (PPO) to further explore global optimal actions.\nThen, a value network is used to run Monte-Carlo tree search (MCTS) to enhance\nthe policy network. We evaluate the performance of our method on four different\ncategories of combinatorial optimization problems and show that our approach\nperforms strongly compared to the state-of-the-art machine learning and\nheuristics based methods.\n", "date": "2022-06-14", "categories": []} {"id": "2206.05990", "title": "Multi-Agent Neural Rewriter for Vehicle Routing with Limited Disclosure\n of Costs", "abstract": " We interpret solving the multi-vehicle routing problem as a team Markov game\nwith partially observable costs. For a given set of customers to serve, the\nplaying agents (vehicles) have the common goal to determine the team-optimal\nagent routes with minimal total cost. Each agent thereby observes only its own\ncost. Our multi-agent reinforcement learning approach, the so-called\nmulti-agent Neural Rewriter, builds on the single-agent Neural Rewriter to\nsolve the problem by iteratively rewriting solutions. Parallel agent action\nexecution and partial observability require new rewriting rules for the game.\nWe propose the introduction of a so-called pool in the system which serves as a\ncollection point for unvisited nodes. It enables agents to act simultaneously\nand exchange nodes in a conflict-free manner. We realize limited disclosure of\nagent-specific costs by only sharing them during learning. During inference,\neach agents acts decentrally, solely based on its own cost. First empirical\nresults on small problem sizes demonstrate that we reach a performance close to\nthe employed OR-Tools benchmark which operates in the perfect cost information\nsetting.\n", "date": "2022-06-13", "categories": []} {"id": "2206.04249", "title": "An Optimization Method-Assisted Ensemble Deep Reinforcement Learning\n Algorithm to Solve Unit Commitment Problems", "abstract": " Unit commitment (UC) is a fundamental problem in the day-ahead electricity\nmarket, and it is critical to solve UC problems efficiently. Mathematical\noptimization techniques like dynamic programming, Lagrangian relaxation, and\nmixed-integer quadratic programming (MIQP) are commonly adopted for UC\nproblems. However, the calculation time of these methods increases at an\nexponential rate with the amount of generators and energy resources, which is\nstill the main bottleneck in industry. Recent advances in artificial\nintelligence have demonstrated the capability of reinforcement learning (RL) to\nsolve UC problems. Unfortunately, the existing research on solving UC problems\nwith RL suffers from the curse of dimensionality when the size of UC problems\ngrows. To deal with these problems, we propose an optimization method-assisted\nensemble deep reinforcement learning algorithm, where UC problems are\nformulated as a Markov Decision Process (MDP) and solved by multi-step deep\nQ-learning in an ensemble framework. The proposed algorithm establishes a\ncandidate action set by solving tailored optimization problems to ensure a\nrelatively high performance and the satisfaction of operational constraints.\nNumerical studies on IEEE 118 and 300-bus systems show that our algorithm\noutperforms the baseline RL algorithm and MIQP. Furthermore, the proposed\nalgorithm shows strong generalization capacity under unforeseen operational\nconditions.\n", "date": "2022-06-09", "categories": []} {"id": "2206.03703", "title": "Sampling-based techniques for designing school boundaries", "abstract": " Recently, an increasing number of researchers, especially in the realm of\npolitical redistricting, have proposed sampling-based techniques to generate a\nsubset of plans from the vast space of districting plans. These techniques have\nbeen increasingly adopted by U.S. courts of law and independent commissions as\na tool for identifying partisan gerrymanders. Motivated by these recent\ndevelopments, we develop a set of similar sampling techniques for designing\nschool boundaries based on the flip proposal. Note that the flip proposal here\nrefers to the change in the districting plan by a single assignment. These\nsampling-based techniques serve a dual purpose. They can be used as a baseline\nfor comparing redistricting algorithms based on local search. Additionally,\nthese techniques can help to infer the problem characteristics that may be\nfurther used for developing efficient redistricting methods. We empirically\ntouch on both these aspects in regards to the problem of school redistricting.\n", "date": "2022-06-08", "categories": []} {"id": "2206.03185", "title": "A new Hyper-heuristic based on Adaptive Simulated Annealing and\n Reinforcement Learning for the Capacitated Electric Vehicle Routing Problem", "abstract": " Electric vehicles (EVs) have been adopted in urban areas to reduce\nenvironmental pollution and global warming as a result of the increasing number\nof freight vehicles. However, there are still deficiencies in routing the\ntrajectories of last-mile logistics that continue to impact social and economic\nsustainability. For that reason, in this paper, a hyper-heuristic (HH) approach\ncalled Hyper-heuristic Adaptive Simulated Annealing with Reinforcement Learning\n(HHASA$_{RL}$) is proposed. It is composed of a multi-armed bandit method and\nthe self-adaptive Simulated Annealing (SA) metaheuristic algorithm for solving\nthe problem called Capacitated Electric Vehicle Routing Problem (CEVRP). Due to\nthe limited number of charging stations and the travel range of EVs, the EVs\nmust require battery recharging moments in advance and reduce travel times and\ncosts. The HH implemented improves multiple minimum best-known solutions and\nobtains the best mean values for some high-dimensional instances for the\nproposed benchmark for the IEEE WCCI2020 competition.\n", "date": "2022-06-07", "categories": []} {"id": "2206.02771", "title": "Neuro CROSS exchange: Learning to CROSS exchange to solve realistic\n vehicle routing problems", "abstract": " CROSS exchange (CE), a meta-heuristic that solves various vehicle routing\nproblems (VRPs), improves the solutions of VRPs by swapping the sub-tours of\nthe vehicles. Inspired by CE, we propose Neuro CE (NCE), a fundamental operator\nof learned meta-heuristic, to solve various VRPs while overcoming the\nlimitations of CE (i.e., the expensive $\\mathcal{O}(n^4)$ search cost). NCE\nemploys a graph neural network to predict the cost-decrements (i.e., results of\nCE searches) and utilizes the predicted cost-decrements as guidance for search\nto decrease the search cost to $\\mathcal{O}(n^2)$. As the learning objective of\nNCE is to predict the cost-decrement, the training can be simply done in a\nsupervised fashion, whose training samples can be prepared effortlessly.\nDespite the simplicity of NCE, numerical results show that the NCE trained with\nflexible multi-depot VRP (FMDVRP) outperforms the meta-heuristic baselines.\nMore importantly, it significantly outperforms the neural baselines when\nsolving distinctive special cases of FMDVRP (e.g., MDVRP, mTSP, CVRP) without\nadditional training.\n", "date": "2022-06-06", "categories": []} {"id": "2606.21604", "title": "Learning to Place Guards by Reinforcement: A Geo-Free Neural Policy for the Vertex-Guard Art Gallery Problem", "abstract": "Neural combinatorial optimization (NCO) has shown that policies trained by reinforcement can construct strong solutions to NP-hard problems directly from raw instances. What such a policy actually learns, as opposed to what its decoder expresses, remains much less clear. We study this distinction on the vertex-guard Art Gallery Problem, the NP-hard task of choosing polygon vertices from which to observe an entire region. A pointer-network policy is trained from a coverage-aware reward over its own rollouts under the constraint we call geo-free inference: at test time it sees only vertex coordinates, with no visibility computation and no geometric oracle. The policy places guards economically but leaves a tail of under-covered polygons that widens far beyond the training range. To locate the cause, we freeze the trained encoder and read its embeddings with a small single-shot classifier, still geo-free at inference. The classifier closes most of the feasibility gap, in and out of distribution and at up to roughly five times the training range, cutting under-covered polygons by about an order of magnitude at an explicitly reported cost in guard count. We read this as evidence that the reinforcement-trained representation already encodes the geometry required for feasibility, and that residual failures reflect decoder calibration rather than missing knowledge. Probing a frozen encoder thus offers a practical way to ask what a neural combinatorial solver has internalized.", "date": "2026-06-19", "categories": []} {"id": "2606.22159", "title": "Deep RL for Fast Long-Horizon Operations Scheduling on NASA's Carruthers Geocorona Observatory Mission", "abstract": "Spacecraft operations scheduling is a highly constrained, long-horizon combinatorial optimization problem that traditionally relies on heuristics, constraint programming, or manual planning. We present a scalable deep reinforcement learning framework developed and deployed for NASA's Carruthers Geocorona Observatory mission. Our framework introduces a macro-action abstraction known as activity blocks coupled with dynamic action-masking to navigate the intractably large search space and strictly enforce complex power, thermal, and instrument constraints. The resulting architecture generates globally feasible schedules with overwhelming probability, establishes operational trust, and executes a full training cycle in under six hours, circumventing the need for policy robustness by enabling rapid, on-demand retraining. Further, resulting schedules outperform baseline heuristics in scheduled science quality. The deep reinforcement learning framework was deployed as the default operational scheduler for the Carruthers Geocorona Observatory mission from the outset of the mission, demonstrating that deep reinforcement learning can be trusted for real spacecraft operations under complex, evolving constraints.", "date": "2026-06-20", "categories": []} {"id": "2606.22776", "title": "GeoRouteNet: Geometry-Enhanced Non-Autoregressive Neural Solver for the Traveling Salesman Problem", "abstract": "The traveling salesman problem (TSP) is a canonical NP-hard combinatorial optimization benchmark that tests the representational capacity and generalization of neural solvers. While non-autoregressive (NAR) approaches offer parallel inference, they often lack sufficient geometric inductive bias and stable training signals, leading to degraded performance under cross-scale and cross-distribution shifts. We propose GeoRouteNet, a geometry-enhanced NAR neural solver for Euclidean TSP. On the model side, GeoRouteNet incorporates centered node features, learnable radial distance basis functions, distance-aware graph attention with explicit edge messaging, LayerNorm-SwiGLU feed-forward blocks, and cross-layer attentive residual mixing. On the training side, we design multi-candidate self-comparison reinforcement learning (MCS-RL), which samples multiple candidate tours per instance, constructs adaptive baselines from greedy and peer candidates, and adds winner-candidate guidance with annealed entropy regularization. On 10,000 random TSP50 instances, GeoRouteNet achieves a 0.32% optimality gap under Beam-1000 decoding. On TSP100, the gap is 1.26%. On 27 stratified TSPLIB EUC_2D instances, the overall gap drops from 17.12% (NAR4TSP reproduction) to 3.60%, while batch inference throughput substantially exceeds that of Concorde and LKH3. Ablation studies confirm that geometric structure enhancement and multi-candidate training are complementary: structure improvements dominate cross-distribution gains, while MCS-RL further stabilizes solution quality when paired with a strong geometric encoder.", "date": "2026-06-22", "categories": []} {"id": "2606.22831", "title": "Learning-Augmented Algorithms for Online Vertex Cover", "abstract": "This paper studies learning-augmented online weighted vertex cover with advice and a parameter $\\lambda \\in (0,1)$. We consider two graph cases: bipartite graphs and general graphs. In both settings, the online algorithm must maintain a feasible vertex cover under irrevocable decisions. We show that these problems admit the same robustness--consistency tradeoffs as learning-augmented ski rental. For the bipartite graph model, we give a randomized algorithm that is $\\frac{1}{1-e^{-\\lambda}}$-robust and $\\frac{\\lambda}{1-e^{-\\lambda}}$-consistent. For the general graph model, we give a deterministic algorithm that is $(1+\\frac{1}{\\lambda})$-robust and $(1+\\lambda)$-consistent. We prove that the tradeoffs above are optimal in both settings. We also validate the proposed algorithms through experiments on synthetic and real-world datasets.", "date": "2026-06-22", "categories": []} {"id": "2606.22978", "title": "Joint Air Traffic Flow and Capacity Management via Answer Set Programming", "abstract": "Operational Air Traffic Flow and Capacity Management (ATFCM) balances flight demand with available sector capacity, to ensure safe and efficient operations. Mathematical models enhance operational ATFCM performance by framing demand-capacity balancing as an optimization problem, maximizing efficiency while adhering to safety constraints. However, SOTA research optimizes the aircraft trajectories (called ATFM) or the sector configuration (called DAC) separately. This leaves a research gap of whether joint optimization of ATFM and DAC can bring benefits. We partially address this limitation by introducing a joint ATFCM model with an encoding in Answer Set Programming (ASP). The ASP implementation is evaluated against two baselines applied to our joint model: a SOTA Mixed Integer Programming (MIP) model and an iterative CASA-based heuristic. Computational experiments utilize an instance generator fitted to historical OpenSky Network flight data. Our results indicate that the ASP model outperforms the MIP model, while ASP remains competitive against heuristics on small instances. Furthermore, while DAC has the largest improvement on solving performance compared to rerouting and delaying, unrestricted variants of DAC or rerouting lead to search space thrashing.", "date": "2026-06-22", "categories": []} {"id": "2606.24672", "title": "Cost-Optimal Decision Diagrams for Stochastic Boolean Function Evaluation", "abstract": "In many decision-making scenarios, acquiring information incurs different costs. We consider the problem of constructing a deterministic evaluation strategy that minimizes the expected cost of evaluating a propositional formula under variable costs and a probability distribution over truth assignments. We present a branch-and-bound algorithm with variable-selection heuristics, pruning, and caching. To the best of our knowledge, it is the first practical exact algorithm for this level of generality. Experiments on random instances demonstrate scalability and quantify the efficiency-quality trade-off of a greedy beam-search variant. We additionally evaluate a structured heart-disease diagnosis instance. Finally, we prove that the problem is $\\#P$-hard and contained in $\\mathrm{PSPACE}$.", "date": "2026-06-23", "categories": []} {"id": "2606.25394", "title": "FactorLibrary: From Polynomials to Circuits via Recursive Subgoals", "abstract": "Finding minimal arithmetic circuits for polynomials over finite fields is a combinatorially hard problem central to algebraic complexity theory. We formulate it as a reinforcement learning problem in two directions, bottom-up and top-down. To address the challenge of a fast-growing combinatorial search space, we introduce FactorLibrary, which stores factorizable subexpressions that serve as reusable subgoals across training episodes. We trained a bottom-up agent with Gumbel-PPO-MCTS and two top-down agents with PPO+MCTS and SAC. The PPO+MCTS top-down agent exhibited the most stable performance, finding certified optimal circuits up to complexity $8$ with a success rate of $91.8\\%$.", "date": "2026-06-24", "categories": []} {"id": "2606.26399", "title": "Geometry-Aware MCTS for Extremal Problems in Combinatorial Geometry", "abstract": "We study certain extremal problems in combinatorial geometry that ask about configurations of points in an $n \\times n$ grid that satisfy strict, global geometric constraints. Classical exact solvers suffer from combinatorial explosion for these types of problems, and standard reinforcement learning and transformer-based models struggle with the sparse reward \"validity cliff\" and quadratic token-consumption limits. To overcome these bottlenecks, we propose a Geometry-Aware Monte Carlo Tree Search (MCTS) framework. Our approach strictly enforces geometric constraints through incremental updates to the feasible action space. For constraints about collections of collinear points, like those that occur in the classic No-Three-in-Line problem (Max-N3IL), this mechanism reduces the constraint checking complexity from $O(n^3)$ to $O(n^2)$. To improve search efficiency, we exploit geometric symmetries in two ways: canonical pruning during node expansion to reduce the branching factor, and symmetric batch transitions to accelerate the discovery of promising configurations. We perform extensive experiments and establish new best-known computational results on five out of six of the problems that we considered. Notably, for Max-N3IL we find configurations of size roughly $1.8 n$ for grids of size $82 \\le n \\le 119$. For the Smallest Complete Set problem, we find configurations of size roughly $0.95 n$, providing new upper bounds within the tested grids. This work establishes Geometry-Aware MCTS as a highly adaptable framework for discovering novel configurations in combinatorial geometry.", "date": "2026-06-24", "categories": []} {"id": "2606.26852", "title": "Context-Aware Synthesis of Optimization Pipelines for Warehouse Optimization", "abstract": "Order fulfillment in manual picker-to-goods warehouses involves interconnected decisions such as item assignment, order batching, and picker routing. While integrated models capture interactions between these decisions, practical warehouse systems often require decomposed approaches due to organizational boundaries, differing responsibilities, or limited data availability. Existing studies primarily evaluate algorithms for isolated subproblems or fixed subproblem combinations for specific warehouse settings, but lack a general mechanism to determine applicable algorithm configurations, compose them into valid solution pipelines, and assess their performance.\n With Context-Aware Synthesis of Optimization Pipelines (CASOP), we propose a framework for constructing and evaluating context-specific optimization pipelines and apply these to order fulfillment. The framework comprises: (1) a modular repository of algorithms for common order fulfillment problems; (2) semantic data and algorithm cards to describe warehouse context and algorithm requirements; (3) a taxonomy that structures order fulfillment problems into relevant subproblems; (4) a pipeline synthesizer that identifies applicable algorithms for a given warehouse context and composes all valid optimization pipelines; and (5) a pipeline evaluator that assesses all resulting pipelines. We demonstrate the framework on 7 benchmark instance sets covering four problem classes, resulting in 1,063,044 valid pipelines. The framework supports researchers and practitioners in designing, automatically synthesizing, and selecting valid, high-performing algorithmic pipelines for warehouse operations. The software is open-source and available at https://github.com/kit-dsm/ware_ops_pipes and https://github.com/kit-dsm/ware_ops_algos.\n Keywords: Warehouse optimization, Algorithm selection, Pipeline synthesis, Order fulfillment", "date": "2026-06-25", "categories": []} {"id": "2606.30680", "title": "Locker-based Truck-Drone Routing with Integrated Considerations of Pickups, Deliveries, and No-Fly Zones", "abstract": "Truck-drone delivery is an emerging last-mile logistics mode combining the long-haul capacity of trucks with the flexible service capability of drones. In locker-based operations, smart lockers serve not only as temporary parcel storage facilities but also as automated drone docking and service nodes. These automated nodes support drone takeoff, landing, parcel handover, and battery replacement, thereby significantly extending the service range and operational flexibility of drone-assisted delivery networks. However, practical locker-based delivery systems face complex real-world challenges, requiring the integrated coordination of not only parcel delivery, return pickup, battery-constrained and load-dependent drone flights, but also necessary detours around restricted airspace. To address this practical and multifaceted challenge, this paper introduces a locker-based truck-drone routing problem with integrated considerations of pickups, deliveries, and no-fly zones (LTDRP-PDNF), with the objective of minimizing the total operational cost of a fleet of drone-equipped trucks. We formulate the route construction process as a Markov Decision Process and develop a two-stage deep reinforcement learning-based neural heuristic. The first stage utilizes an attention-based encoder and a Bidirectional Gated Recurrent Unit decoder to solve the truck-only routing problem, formulated as a capacitated vehicle routing problem. The second stage combines a policy-transfer strategy with a hybrid dispatch assignment heuristic to construct fully coordinated truck and drone routes for LTDRP-PDNF. Experiments on instances of different scales demonstrate that the proposed method outperforms metaheuristic and neural heuristic baselines in most cases while maintaining exceptionally short computation times, offering an effective, scalable solution framework under practical operational constraints.", "date": "2026-06-26", "categories": []} {"id": "2606.29725", "title": "Optimizing Nursing Care Taxi Dispatch Leveraging Integer Linear Programming Solvers and Machine Learning", "abstract": "In this paper, we formulate a new vehicle dispatch optimization problem, called Nursing Care Taxi Dispatch, as a variant of the Vehicle Routing Problem, considering constraints related to wheelchair use, user compatibility, pick-up and drop-off times, and vehicle limitations. Previous neural-based methods for Vehicle Routing Problems have typically addressed a few simple constraints, while our new problem involves multiple complex constraints, resulting in having fewer destinations to select. This complexity makes it more difficult to obtain solutions that allow all nodes to be visited with a limited number of vehicles. To balance low violation rate, computational efficiency, and solution quality, we propose a supervised machine learning approach based on the Transformer architecture. We first obtain a set of high-quality solutions using an integer linear programming solver for given inputs and then train our learning model through supervised learning. Additionally, we introduce the post-processing of the paths generated by the learning model, ensuring that all constraints are satisfied. We compared each instance's objective function value (operating time), execution time, and constraint violation rate across different methods: our proposed method and some existing methods including integer linear programming and machine learning-based methods, using real-world facility data. Our method successfully produced balanced solutions regarding operating time, execution time, and constraint violation rate. Notably, we observed a decrease in the operating time for all problem sizes and regions, while keeping constraint violations to a minimum compared to existing methods. Especially, the decrease reached up to 8% for problem sizes with fewer than 30 users.", "date": "2026-06-29", "categories": []} {"id": "2606.30333", "title": "Local-Minima-Preserving Continuous Relaxation of Ising Problems", "abstract": "The generalized Ising problem captures a broad spectrum of hard combinatorial problems, including MAX-CUT, Number Partitioning (NPP), and Maximum Independent Set. In this work, we consider the notion of one-flip local minima for this problem. We construct a polynomial relaxation and prove the landscape equivalence theorem: there exists a one-to-one correspondence between the local minima of the relaxation and the one-flip minima of the original Ising problem. This guarantee reduces the Ising problem to finding the local minima of a smooth function, allowing us to leverage gradient-based optimizers such as ADAM. We demonstrate that our method is scalable and it achieves strong performance across challenging benchmarks, including spin-glass models, MAX-CUT, and NPP.", "date": "2026-06-29", "categories": []} {"id": "2606.31820", "title": "Adaptive Cluster-First Route-Second Decomposition for Industrial-Scale Vehicle Routing", "abstract": "Large-scale capacitated vehicle routing problems (CVRPs) are commonly addressed using cluster-first route-second (CFRS) approaches that split a routing instance into smaller, computationally tractable subproblems. Existing splitting methods typically rely on fixed partitioning rules, predefined optimization objectives, or learned policies, which may perform inconsistently across instances exhibiting different spatial, demand, and operational characteristics. In this work, we propose an adaptive CFRS system that formulates a decomposition procedure as an iterative decision-making process. Motivated by the recent success of large language models (LLMs) in reasoning and tool selection, the system employs an LLM as a high-level decision maker that analyzes the evolving decomposition state and selectively applies further clustering, balancing, and refinement operators. The proposed algorithm jointly partitions customers and vehicles, enabling capacity-aware clustering while adapting partitioning decisions to the characteristics of each problem. We evaluate the approach on synthetic and benchmark-derived CVRP instances containing up to 500,000 customers. Experimental results demonstrate competitive performance on benchmark-scale instances while exhibiting improved scalability and robust routing quality on substantially larger problems. These results highlight the potential of adaptive, LLM-guided decision support as a practical approach for industrial-scale vehicle routing and large-scale logistics planning.", "date": "2026-06-30", "categories": []}